Performance metrics available in the GUI
The performance metrics defined in the performance tool helps to collect the performance data based on various aspects.
The following tables list the available metrics in each resource type and its descriptions. It also provides a mapping between the performance metrics in the GUI with the corresponding mmperfmon metric names.
Resource type - Network
The following table lists the performance metrics available for the resource type - Network.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
Network – Node, Adapter, Adapter per node,
Cluster1 Note: 1: The aggregation levels such as Node and
Cluster consider the network traffic of network interfaces that is caused by the
ESS system. This includes GPFS data traffic, GPFS administration traffic, and CES network traffic.
|
Network | Throughput - Bytes received | netdev_bytes_r | Number of bytes of data received by the network interface. | bytes/s |
Throughput - Bytes sent | netdev_bytes_s | Number of bytes of data sent by the network interface. | bytes/s | ||
I/O - Packets received | netdev_packets_r | Number of packets received by the network interface. | 1/s | ||
I/O - Packets sent | netdev_packets_s | Number of packets sent by the network interface. | 1/s | ||
I/O - Compressed received | netdev_compressed_r | Number of compressed packets received by the device driver of the network interface. | 1/s | ||
I/O - Compressed sent | netdev_compressed_s | Number of compressed packets sent by the device driver of the network interface. | 1/s | ||
I/O - Multicast frames | netdev_multicast_r | Number of multicast frames received by the device driver of the network interface. | 1/s | ||
Errors - Collisions | netdev_collisions | Number of collisions detected on the network interface. | 1/s | ||
Errors - Carrier loss | netdev_carrier | Number of carrier losses detected on the network interface. | 1/s | ||
Errors - Drops received | netdev_drops_r | Number of incoming packets dropped by the network interface. | 1/s | ||
Errors - Drops sent | netdev_drops_s | Number of outgoing packets dropped by the device driver of the network interface. | 1/s | ||
Errors - Errors received | netdev_errors_r | Number of receive errors detected by the network interface. | 1/s | ||
Errors - Errors sent | netdev_errors_s | Number of send errors detected by the network interface. | 1/s | ||
Errors - FIFO receive error | netdev_fifo_r | Number of FIFO buffer errors reported while receiving data. | 1/s | ||
Errors - FIFO transmit error | netdev_fifo_s | Number of FIFO buffer errors reported while sending data. | 1/s | ||
Errors - Frame errors | netdev_frames_r | Number of packet framing errors reported while receiving data. | 1/s |
Resource type - Infiniband Network
The following table lists the performance metrics available for the resource type - Infiniband Network.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
Infiniband Network – Node, Adapter, Adapter per node, Cluster | Infiniband | Transmitted Data | port_xmit_data | Total number of data octets, which are divided by 4 (lanes), transmitted through all virtual lanes. | 1/s |
Received Data | port_rcv_data | Total number of data octets, which are divided by 4 (lanes), received on all virtual lanes. | 1/s | ||
Transmitted Packets | port_xmit_packets | Total number of packets that are transmitted through all virtual lanes from the port. This might include erroneous packets as well. | 1/s | ||
Received Packets | port_rcv_packets | Total number of packets that are received through all virtual lanes. This might include erroneous packets as well. | 1/s | ||
Received Packets with Errors | port_rcv_errors | Total number of erroneous packets that are received on the port. These errors include local physical errors an all physical errors, malformed data packet errors, and packets discarded due to buffer overrun. | 1/s |
Resource type - System Resources
The following table lists the performance metrics available for the resource type - System Resources.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
System Resources – Node,Cluster | CPU Load Memory |
Rates - Context switches | cpu_contexts | Number of context switches across all CPU cores. | 1/s |
Usage - Nice | cpu_nice | Percentage of total CPU time spent in lowest-priority user processes. | % | ||
Usage - System | cpu_system | Percentage of CPU cycles used by the kernel. | % | ||
General - Inactive | mem_inactive | Inactive memory that is not accessed recently. | MiB | ||
Interrupts - Hardware | cpu_hiq | Percentage of total CPU cycles spent in serving hardware interrupts. | % | ||
General - Active anon | mem_active_anon | Active memory with no file association. That is, heap and stack memory. | MiB | ||
General - Cached | mem_cached | In-memory cache for file reads from the disk. Also known as page cache. This cache does not include the cache for the pages that are swapped back. | MiB | ||
General - Swap cached | mem_swapcached | In-memory cache for pages that are swapped back. | MiB | ||
General - Free | mem_memfree | Unused memory (RAM) in the system. | MiB | ||
General - Inactive file | mem_inactive_file | Inactive memory that is associated with a file, for example, page cache memory. | MiB | ||
General - Active | mem_active | Memory that is being used by a an active process. | MiB | ||
Rates - Interrupts | cpu_interrupts | Number of serviced CPU interrupts. | 1/s | ||
General - Used | mem_memfree / mem_memtotal | Percentage of the used memory out of available memory. | % | ||
General - Buffers | mem_buffers | Memory used for critical system buffers. Buffer memory is the temporary storage used for raw disk blocks. | MiB | ||
Load - Load | load1 | The average number of processes in the run queue over the last 1 minute. | This is just a number, so no unit | ||
General - Active file | mem_active_file | Active memory that is associated with a file, for example, page cache memory. | MiB | ||
Usage - User | cpu_user | Percentage of total CPU time spent in normal priority user processes. | % | ||
Usage - Usage | cpu_idle | Percentage of CPU capacity used by all processes. | % | ||
Interrupts - Software | cpu_siq | Percentage of total CPU cycles spent in serving software interrupts. | % | ||
Usage - I/O wait | cpu_iowait | Percentage of total CPU cycles spent in waiting for I/O to complete. | % | ||
Usage - Idle | cpu_idle | Percentage of CPU cycles that is left unused. | % | ||
General - Inactive anon | mem_inactive_anon | Inactive heap and stack memory with no file association. | MiB |
Resource type - Native RAID
The following table lists the performance metrics available for the resource type - Native RAID.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
Native RAID - pdisk | GPFSPDDisk | Throughput - Bytes read | gpfs_pdds_bytes_read | Number of bytes of data read from the Pdisks. | bytes/s |
Throughput - Bytes written | gpfs_pdds_bytes_written | Number of bytes of data written to the Pdisks. | bytes/s | ||
Throughput - Read ops | gpfs_pdds_read_ops | Number of read operations from Pdisks. | ops/s | ||
Throughput - Write ops | gpfs_pdds_write_ops | Number of write operations on Pdisks. | ops/s | ||
Latency - Avg disk wait r | gpfs_pdds_tot_disk_wait_rd / gpfs_pdds_read_ops | Average Pdisk wait time per read operation. | s (ns, ms, us) | ||
Latency - Avg disk wait w | gpfs_pdds_tot_disk_wait_wr / gpfs_pdds_write_ops | Average Pdisk wait time per write operation. | s (ns, ms, us) | ||
Latency - Min disk wait r | gpfs_pdds_min_disk_wait_rd | Minimum Pdisk wait time per read operation. | s (ns, ms, us) | ||
Latency - Min disk wait w | gpfs_pdds_min_disk_wait_wr | Minimum Pdisk wait time per write operation. | s (ns, ms, us) | ||
Latency - Max disk wait r | gpfs_pdds_max_disk_wait_rd | Maximum Pdisk wait time per read operation. | s (ns, ms, us) | ||
Latency - Max disk wait w | gpfs_pdds_max_disk_wait_wr | Maximum Pdisk wait time per write operation. | s (ns, ms, us) | ||
Latency - Avg queue wait r | gpfs_pdds_tot_queue_wait_rd / gpfs_pdds_read_ops | Average queue wait per read operation on Pdisks. | s (ns, ms, us) | ||
Latency - Avg queue wait w | gpfs_pdds_tot_queue_wait_wr / gpfs_pdds_write_ops | Average queue wait per write operation on Pdisks. | s (ns, ms, us) | ||
Latency - Min queue wait r | gpfs_pdds_min_queue_wait_rd | Minimum queue wait per read operation on Pdisks. | s (ns, ms, us) | ||
Latency - Min queue wait w | gpfs_pdds_min_queue_wait_wr | Minimum queue wait per write operation on Pdisks. | s (ns, ms, us) | ||
Latency - Max queue wait r | gpfs_pdds_max_queue_wait_rd | Maximum queue wait per read operation on Pdisks. | s (ns, ms, us) | ||
Latency - Max queue wait w | gpfs_pdds_max_queue_wait_wr | Maximum queue wait per write operation on Pdisks. | s (ns, ms, us) |
Resource type - NSD Server
The following table lists the performance metrics available for the resource type - NSD Server.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
NSD Server – NSD, Node, Cluster | GPFSNSDDisk | Throughput - Bytes read | gpfs_nsdds_bytes_read | Number of bytes of data reads from all the NSD disks that are attached to the node. | bytes/s |
Throughput - Bytes written | gpfs_nsdds_bytes_written | Number of bytes of data written to all the NSD disks that are attached to the node. | bytes/s | ||
Throughput - Read ops | gpfs_nsdds_read_ops | Number of read operations from all NSD disks. | ops/s | ||
Throughput - Write ops | gpfs_nsdds_write_ops | Number of write operations to all NSD disks. | ops/s | ||
Latency - Avg disk wait r | gpfs_nsdds_tot_disk_wait_rd / gpfs_nsdds_read_ops | Average NSD disk wait time per read operation. | s (ns, ms, us) | ||
Latency - Avg disk wait w | gpfs_nsdds_tot_disk_wait_wr / gpfs_nsdds_write_ops | Average NSD disk wait time per write operation. | s (ns, ms, us) | ||
Latency - Min disk wait r | gpfs_nsdds_min_disk_wait_rd | The shortest time spent waiting for an NSD disk read operation. | s (ns, ms, us) | ||
Latency - Min disk wait w | gpfs_nsdds_min_disk_wait_wr | The shortest time spent waiting for a write operation on the NSD disk. | s (ns, ms, us) | ||
Latency - Max disk wait r | gpfs_nsdds_max_disk_wait_rd | The longest time between an NSD disk read operation initialization and the completion of that operation. | s (ns, ms, us) | ||
Latency - Max disk wait w | gpfs_nsdds_max_disk_wait_wr | The longest time between an NSD disk read operation initialization and the completion of that operation. | s (ns, ms, us) | ||
Latency - Avg queue wait r | gpfs_nsdds_tot_queue_wait_rd / gpfs_nsdds_read_ops | Average queue wait per read operation from NSD disks. | s (ns, ms, us) | ||
Latency - Avg queue wait w | gpfs_nsdds_tot_queue_wait_wr / gpfs_nsdds_write_ops | Average queue wait per write operation to the NSD disks. | s (ns, ms, us) | ||
Latency - Min queue wait r | gpfs_nsdds_min_queue_wait_rd | Minimum queue wait per read operation from the NSD disks. | s (ns, ms, us) | ||
Latency - Min queue wait w | gpfs_nsdds_min_queue_wait_wr | Minimum queue wait per write operation to the NSD disks. | s (ns, ms, us) | ||
Latency - Max queue wait r | gpfs_nsdds_max_queue_wait_rd | Maximum queue wait per read operation from the NSD disks. | s (ns, ms, us) | ||
Latency - Max queue wait w | gpfs_nsdds_max_queue_wait_wr | Maximum queue wait per write operation to the NSD disks. | s (ns, ms, us) | ||
NSD Server - File System, Pool | Throughput - Bytes read | gpfs_nsdfs_bytes_read | Number of bytes read in the NSD disks that are aggregated to the file system. | bytes/s | |
Throughput - Bytes written | gpfs_nsdfs_bytes_written | Number bytes written in the NSD disks that are aggregated to the file system. | bytes/s | ||
Throughput - Read ops | gpfs_nsdfs_read_ops | Number of read operations in the NSD disks that are aggregated to the file system. | ops/s | ||
Throughput - Write ops | gpfs_nsdfs_write_ops | Number of write operations in the NSD disks that are aggregated to the file system. | ops/s |
Resource type - IBM Spectrum Scale Client
The following table lists the performance metrics available for the resource type - IBM Spectrum Scale Client.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
IBM Spectrum Scale™ Client - File System | GPFSFilesystem GPFSVFS GPFSFilesystemAPI Note: Starting with IBM Spectrum
Scale 5.1.0, the GPFSVFSX sensor replaces the GPFSVFS
sensor and it is enabled by default on new installations. When the GPFSVFSX sensor is enabled, the
GUI uses the gpfs_vfsx_ metrics instead of the gpfs_vfs_ metrics. For example, gpfs_vfsx_create is
used instead of gpfs_vfs_create.
|
Throughput - Bytes read | gpfs_fis_bytes_read | Total number of bytes read from both disk and cache. | bytes/s |
Throughput - Bytes written | gpfs_fis_bytes_written | Total number of bytes written to both disk and cache. | bytes/s | ||
Throughput - Read ops | gpfs_fis_read_calls | Number of read operations on all mounted file systems of IBM Spectrum Scale client. | ops/s | ||
Throughput - Write ops | gpfs_fis_write_calls | Number of write operations on all mounted file systems of IBM Spectrum Scale client. | ops/s | ||
Latency - Avg disk wait r | gpfs_fs_tot_disk_wait_rd / gpfs_fs_read_ops | Average disk wait time per read operation on the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Avg disk wait w | gpfs_fs_tot_disk_wait_wr / gpfs_fs_write_ops | Average disk wait time per write operation on the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min disk wait r | gpfs_fs_min_disk_wait_rd | The shortest time spent in waiting for a disk read operation in the mounted file systems of IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min disk wait w | gpfs_fs_min_disk_wait_wr | The shortest time spent in waiting for a write operation in the mounted file systems of IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Max disk wait r | gpfs_fs_max_disk_wait_rd | The longest time between a read operation initialization and the completion of that operation in the file systems of IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Max disk wait w | gpfs_fs_max_disk_wait_wr | The longest time between a write operation initialization and the completion of that operation in the mounted file systems of IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Avg queue wait r | gpfs_fs_tot_queue_wait_rd / gpfs_fs_read_ops | Average queue wait per read operation on the mounted file systems of IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Avg queue wait w | gpfs_fs_tot_queue_wait_wr / gpfs_fs_write_ops | Average queue wait per write operation on the mounted file system of IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min queue wait r | gpfs_fs_min_queue_wait_rd | The shortest time a disk read operation is queued to complete that operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min queue wait w | gpfs_fs_min_queue_wait_wr | The shortest time a disk write operation is queued to complete that operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Max queue wait r | gpfs_fs_max_queue_wait_rd | The longest time a disk read operation is queued to complete that operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Max queue wait w | gpfs_fs_max_queue_wait_wr | The longest time a disk write operation is queued to complete that operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
IBM Spectrum Scale™ Client - Nodes | Latency - Rename | gpfs_vfs_rename_t / gpfs_vfs_rename | Average latency of rename operations per node in the IBM Spectrum Scale client. | s (ns, ms, us) | |
Operations - Readdir | gpfs_fis_readdir_calls | Number of readdir operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Link | gpfs_vfs_link | Number of link operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Avg disk wait w | gpfs_fs_tot_disk_wait_wr / gpfs_fs_write_ops | Average disk wait time per write operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Write | gpfs_vfs_write | Number of write operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Listxattr | gpfs_vfs_listxattr | Number of listxattr operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Setattr | gpfs_vfs_setattr | Number of setattr operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Mknod | gpfs_vfs_mknod_t / gpfs_vfs_mknod | Average latency of mknod operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Tsfsattr | gpfs_vfs_tsfsattr | Number of tsfsattr operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Remove | gpfs_vfs_remove | Number of remove operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Mmapwrite | gpfs_vfs_mmapwrite | Number of mmapwrite operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Tsfsattr | gpfs_vfs_tsfsattr_t/gpfs_vfs_tsfsattr | Average latency of tsfsattr operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - GetParent | gpfs_vfs_getParent | Number of getParent operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Max queue wait r | gpfs_fs_max_queue_wait_rd | Maximum queue wait per read operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Symlink | gpfs_vfs_symlink | Number of symlink operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Readlink | gpfs_vfs_readlink_t / gpfs_vfs_readlink | Average latency of readlink operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Lockctl | gpfs_vfs_lockctl_t / gpfs_vfs_lockctl | Average latency of lockctl operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Getxattr | gpfs_vfs_getxattr_t / gpfs_vfs_getxattr | Average latency of getxattr operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min queue wait w | gpfs_fs_min_queue_wait_wr | Minimum queue wait per write operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Ftrunc | gpfs_vfs_ftrunc_t / gpfs_vfs_ftrunc | Average latency of ftrunc operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Fsync | gpfs_vfs_fsync | Number of fsync operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Sync | gpfs_vfs_sync_t / gpfs_vfs_sync | Average latency of sync operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Mkdir | gpfs_vfs_mkdir_t / gpfs_vfs_mkdir | Average latency of mkdir operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - GetDentry | gpfs_vfs_getDentry_t / gpfs_vfs_getDentry | Average latency of getDentry operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Aiowrite | gpfs_vfs_aiowrite | Number of aiowrite operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Getattr | gpfs_vfs_getattr_t / gpfs_vfs_getattr | Average latency of getattr operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Read | gpfs_vfs_read | Number of read operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Max disk wait r | gpfs_fs_max_disk_wait_rd | Maximum disk wait time per read operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Read | gpfs_vfs_read_t / gpfs_vfs_read | Average latency of read operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Create | gpfs_vfs_create | Number of create operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - MapLloff | gpfs_vfs_mapLloff | Number of mapLloff operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Mount | gpfs_vfs_mount_t / gpfs_vfs_mount | Average latency of mount operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Close | gpfs_vfs_close_t / gpfs_vfs_close | Average latency of close operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Aiowrite | gpfs_vfs_aiowrite_t / gpfs_vfs_aiowrite | Average latency of aiowrite operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Readlink | gpfs_vfs_readlink | Number of readlink operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Statfs | gpfs_vfs_statfs | Number of statfs operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - GetParent | gpfs_vfs_getParent_t / gpfs_vfs_getParent | Average latency of getParent operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Flock | gpfs_vfs_flock_t / gpfs_vfs_flock | Average latency of flock operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Lockctl | gpfs_vfs_lockctl | Number of lockctl operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Write | gpfs_vfs_write_t / gpfs_vfs_write | Average latency of write operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Aioread | gpfs_vfs_aioread_t / gpfs_vfs_toreador | Average latency of aioread operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Tsfattr | gpfs_vfs_tsfattr_t / gpfs_vfs_tsfattr | Average latency of tsfattr operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Statfs | gpfs_vfs_statfs_t / gpfs_vfs_statfs | Average latency of statfs operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Readpage | gpfs_vfs_readpage | Number of readpage operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Rmdir | gpfs_vfs_rmdir | Number of rmdir operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Close | gpfs_fis_close_calls | Number of close() call requests serviced by the IBM Spectrum Scale client. | ops/s | ||
Operations - Fsync range | gpfs_vfs_fsyncRange | Number of fsync range operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - EncodeFh | gpfs_vfs_encodeFh_t / gpfs_vfs_encodeFh | Average latency of encodeFh operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Create | gpfs_vfs_create_t / gpfs_vfs_create | Average latency of create operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Rename | gpfs_vfs_rename | Number of rename operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Tsfattr | gpfs_vfs_tsfattr | Number of tsfattr operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Symlink | gpfs_vfs_symlink_t / gpfs_vfs_symlink | Average latency of symlink operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Avg queue wait w | gpfs_fs_tot_queue_wait_wr / gpfs_fs_write_ops | Average queue wait per write operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Getxattr | gpfs_vfs_getxattr | Number of getxattr operations in the IBM Spectrum Scale client. | ops/s | ||
Throughput - Write ops | gpfs_fs_write_ops | Number of write operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Fsync | gpfs_vfs_fsync_t / gpfs_vfs_fsync | Average latency of fsync operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Setattr | gpfs_vfs_setattr_t / gpfs_vfs_setattr | Average latency of setattr operations in the IBM Spectrum Scale client. | s (ns, ms, us | ||
Operations - Getattr | gpfs_vfs_getattr | Number of getattr operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Accesses | gpfs_vfs_accesses | Number of accesses operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Sync | gpfs_vfs_sync | Number of sync operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Lookup | gpfs_vfs_lookup_t / gpfs_vfs_lookup | Average latency of lookup operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Open | Count of open() call requests serviced by the IBM Spectrum Scale client. | ops/s | |||
Operations - Setxattr | gpfs_vfs_setxattr | Number of setxattr operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Setacl | gpfs_vfs_setacl | Number of setacl operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Max queue wait w | gpfs_fs_max_queue_wait_wr | Maximum queue wait per write operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Unmap | gpfs_vfs_unmap | Number of unmap operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Mkdir | gpfs_vfs_mkdir | Number of mkdir operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - MapLloff | gpfs_vfs_mapLloff_t / gpfs_vfs_mapLloff | Average latency of mapLloff operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Readdir | gpfs_vfs_readdir_t / gpfs_vfs_readdir | Average latency of readdir operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Vget | gpfs_vfs_vget_t / gpfs_vfs_vget | Average latency of vget operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min disk wait r | gpfs_fs_min_disk_wait_rd | Minimum disk wait time per read operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Throughput - Bytes read | gpfs_fs_bytes_read | Bytes of data read in the IBM Spectrum Scale client. | bytes/s | ||
Operations - Clear | gpfs_vfs_clear | Number of clear operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Mknod | gpfs_vfs_mknod | Number of mknod operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Writepage | gpfs_vfs_writepage_t / gpfs_vfs_writepage | Average latency of writepage operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - GetDentry | gpfs_vfs_getDentry | Number of getDentry operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Mmapread | gpfs_vfs_mmapread | Number of mmapread operations in the IBM Spectrum Scale client. | ops/s | ||
Throughput - Read ops | gpfs_fs_read_ops | Number of read operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Writepage | gpfs_vfs_writepage | Number of writepage operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Removexattr | gpfs_vfs_removexattr | Number of removexattr operations in the IBM Spectrum Scale client. | ops/s | ||
Throughput - Bytes written | gpfs_fs_bytes_written | Bytes of data written in the IBM Spectrum Scale client. | bytes/s | ||
Operations - Lookup | gpfs_vfs_lookup | Number of lookup operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Mmapread | gpfs_vfs_mmapread_t / gpfs_vfs_mmapread | Average latency of mmapread operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Clear | gpfs_vfs_clear_t / gpfs_vfs_clear | Average latency of clear operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Fsync range | gpfs_vfs_fsyncRange_t / gpfs_vfs_fsyncRange | Average latency of fsync range operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Flock | gpfs_vfs_flock | Number of flock operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - DecodeFh | gpfs_vfs_decodeFh | Number of decodeFh operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Ftrunc | gpfs_vfs_ftrunc | Number of ftrunc operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Setxattr | gpfs_vfs_setxattr_t / gpfs_vfs_setxattr | Average latency of setxattr operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - EncodeFh | gpfs_vfs_encodeFh | Number of encodeFh operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Readpage | gpfs_vfs_readpage_t / gpfs_vfs_readpage | Average latency of readpage operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Link | gpfs_vfs_link_t / gpfs_vfs_link | Average latency of link operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Avg queue wait r | gpfs_fs_tot_queue_wait_rd/ gpfs_fs_read_ops | Average queue wait per read operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min queue wait r | gpfs_fs_min_queue_wait_rd | Minimum queue wait per read operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Vget | gpfs_vfs_vget | Number of vget operations in the IBM Spectrum Scale client. | ops/s | ||
Operations - Aioread | gpfs_vfs_aioread | Number of aioread operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Avg disk wait r | gpfs_fs_tot_disk_wait_rd / gpfs_fs_read_ops | Average disk wait time per read operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Max disk wait w | gpfs_fs_max_disk_wait_wr | Maximum disk wait time per write operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Setacl | gpfs_vfs_setacl_t / gpfs_vfs_setacl | Average latency of setacl operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Unmap | gpfs_vfs_unmap_t / gpfs_vfs_unmap | Average latency of unmap operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Rmdir | gpfs_vfs_rmdir_t / gpfs_vfs_rmdir | Average latency of rmdir operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Removexattr | gpfs_vfs_removexattr_t / gpfs_vfs_removexattr | Average latency of removexattr operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Operations - Mount | gpfs_vfs_mount | Number of mount operations in the IBM Spectrum Scale client. | ops/s | ||
Latency - Mmapwrite | gpfs_vfs_mmapwrite_t / gpfs_vfs_mmapwrite | Average latency of mmapwrite operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Open | gpfs_vfs_open / gpfs_vfs_open_t | Average latency of open operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - DecodeFh | gpfs_vfs_decodeFh_t / gpfs_vfs_decodeFh | Average latency of decodeFh operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Remove | gpfs_vfs_remove_t / gpfs_vfs_remove | Average latency of remove operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Listxattr | gpfs_vfs_listxattr_t / gpfs_vfs_listxattr | Average latency of listxattr operations in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Min disk wait w | gpfs_fs_min_disk_wait_wr | Minimum disk wait time per write operation in the IBM Spectrum Scale client. | s (ns, ms, us) | ||
Latency - Accesses | gpfs_vfs_accesses_t / gpfs_vfs_accesses | Average latency of accesses operations in the IBM Spectrum Scale client. | s (ns, ms, us) |
Resource type - NFS
The following table lists the performance metrics available for the resource type - NFS.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
NFS – Node, Cluster, Export | NFSIO | Throughput - Read | nfs_read | Number of bytes transferred in NFS read operations. | bytes/s |
Throughput - Write | nfs_write | Number of bytes transferred in NFS write operations. | bytes/s | ||
Throughput - V3 Read | nfs_read | Number of bytes transferred in NFSv3 read operations. | bytes/s | ||
Throughput - V3 Write | nfs_write | Number of bytes transferred in NFSv3 write operations. | bytes/s | ||
Throughput - V4 Read | nfs_read | Number of bytes transferred in NFSv4 read operations. | bytes/s | ||
Throughput - V4 Write | nfs_write | Number of bytes transferred in NFSv4 write operations. | bytes/s | ||
Latency - Read processing | nfs_read_lat / nfs_read_ops | Time taken for NFS read operations. | s (ns, ms, us) | ||
Latency - Write processing | nfs_write_lat / nfs_write_ops | Time taken for NFS write operations. | s (ns, ms, us) | ||
Latency - V3 Read processing | nfs_read_lat / nfs_read_ops | Time taken for NFSv3 read operations. | s (ns, ms, us) | ||
Latency - V3 Write processing | nfs_write_lat / nfs_write_ops | Time taken for NFSv3 write operations. | s (ns, ms, us) | ||
Latency - V4 Read processing | nfs_read_lat / nfs_read_ops | Time taken for NFSv4 read operations. | s (ns, ms, us) | ||
Latency - V4 Write processing | nfs_write_lat / nfs_write_ops | Time taken for NFSv4 write operations. | s (ns, ms, us) | ||
Latency - Read queue wait | nfs_read_queue / nfs_read_ops | Time spent in the rpc wait queue for the NFS read operations. | s (ns, ms, us) | ||
Latency - Write queue wait | nfs_write_queue / nfs_write_ops | Time spent in the rpc wait queue for the NFS write operations. | s (ns, ms, us) | ||
Latency - V3 Read queue wait | nfs_read_queue / nfs_read_ops | Time spent in the rpc wait queue for the NFSv3 read operations. | s (ns, ms, us) | ||
Latency - V3 Write queue wait | nfs_write_queue / nfs_write_ops | Time spent in the rpc wait queue for the NFSv3 write operations. | s (ns, ms, us) | ||
Latency - V4 Read queue wait | nfs_read_queue / nfs_read_ops | Time spent in the rpc wait queue for the NFSv4 read operations. | s (ns, ms, us) | ||
Latency - V4 Write queue wait | nfs_write_queue / nfs_write_ops | Time spent in the rpc wait queue for the NFSv4 write operations. | s (ns, ms, us) | ||
Operations - Read IOPS | nfs_read_ops | Number of NFS read operations. | ops/s | ||
Operations - Write IOPS | nfs_write_ops | Number of NFS write operations. | ops/s | ||
Operations - V3 Read IOPS | nfs_read_ops | Number of NFSv3 read operations. | ops/s | ||
Operations - V3 Write IOPS | nfs_write_ops | Number of NFSv3 write operations. | ops/s | ||
Operations - V4 Read IOPS | nfs_read_ops | Number of NFSv4 read operations. | ops/s | ||
Operations - V4 Write IOPS | nfs_write_ops | Number of NFSv4 write operations. | ops/s | ||
Operations - Read errors | nfs_read_err | Number of erroneous NFS read operations. | ops/s | ||
Operations - Write errors | nfs_write_err | Number of erroneous NFS write operations. | ops/s | ||
Operations - V3 Read errors | nfs_read_err | Number of erroneous NFSv3 read operations. | ops/s | ||
Operations - V3 Write errors | nfs_write_err | Number of erroneous NFSv3 write operations. | ops/s | ||
Operations - V4 Read errors | nfs_read_err | Number of erroneous NFSv4 read operations. | ops/s | ||
Operations - V4 Write errors | nfs_write_err | Number of erroneous NFSv4 write operations. | ops/s | ||
Throughput - Read requested | nfs_read_req | Number of bytes requested for NFS read operations. | bytes/s | ||
Throughput - Write requested | nfs_write_req | Number of bytes requested for NFS write operations. | bytes/s | ||
Throughput - V3 Read requested | nfs_read_req | Number of bytes requested for NFSv3 read operations. | bytes/s | ||
Throughput - V3 Write requested | nfs_write_req | Number of bytes requested for NFSv3 write operations. | bytes/s | ||
Throughput - V4 Read requested | nfs_read_req | Number of bytes requested for NFSv4 read operations. | bytes/s | ||
Throughput - V4 Write requested | nfs_write_req | Number of bytes requested for NFSv4 write operations. | bytes/s |
Resource type - SMB
The following table lists the performance metrics available for the resource type - SMB.
Resource type and aggregation levels | Sensor names | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
SMB – Node, Cluster | SMBStats SMBGlobalStats |
SMB Overall - Connect | connect_count | Number of connections made since the initialization of parent smbd process. | Number; so no units. |
SMB Overall - Disconnect | disconnect_count | Number of connections closed since the initialization of parent smbd process. | Number; so no units. | ||
SMB Overall - CPU user | cpu_user_time | The percentage of CPU availed by the users. The user time is determined by the get_rusage system calls. | % | ||
SMB Overall - CPU system | cpu_system_time | The percentage of CPU availed for the system operations. The system time is determined by the get_rusage system call. | % | ||
SMB Overall - Request count | request_count | Number of SMB requests since the initialization. | Number; so no units. | ||
Create - Sent | op_outbytes | Number of bytes sent for the SMB create request. The byte count also includes protocol headers. | bytes/s | ||
Create - Received | op_inbytes | Number of bytes received for the SMB create request. The byte count also includes protocol headers. | bytes/s | ||
Create - Operations | op_count | Number of times the corresponding SMB create request is called. | ops/s | ||
Create - Latency | op_time / op_count | Total amount of time spent for all types of SMB create requests. | s (ns, ms, us) | ||
Close - Sent | op_outbytes | Number of bytes sent for the SMB close request. The byte count also includes protocol headers. | bytes/s | ||
Close - Received | op_inbytes | Number of bytes received for the SMB close request. The byte count also includes protocol headers. | bytes/s | ||
Close - Operations | op_count | Number of times the corresponding SMB close request is called. | ops/s | ||
Close - Latency | op_time / op_count | Total amount of time spent for all types of SMB close requests. | s (ns, ms, us) | ||
Flush - Sent | op_outbytes | Number of bytes sent for the SMB flush request. The byte count also includes protocol headers. | bytes/s | ||
Flush - Received | op_inbytes | Number of bytes received for the SMB flush request. The byte count also includes protocol headers. | bytes/s | ||
Flush - Operations | op_count | Number of times the corresponding SMB flush request is called. | ops/s | ||
Flush - Latency | op_time / op_count | Total amount of time spent for all types of SMB flush requests. | s (ns, ms, us) | ||
Read - Sent | op_outbytes | Number of bytes sent for the SMB read requests. The byte count also includes protocol headers. | bytes/s | ||
Read - Received | op_inbytes | Number of bytes received for the SMB read requests. The byte count also includes protocol headers. | bytes/s | ||
Read - Operations | op_count | Number of times the corresponding SMB read request is called. | ops/s | ||
Read - Latency | op_time / op_count | Total amount of time spent for all types of SMB read requests. | s (ns, ms, us) | ||
Write - Sent | op_outbytes | Number of bytes sent for the SMB write requests. The byte count also includes protocol headers. | bytes/s | ||
Write - Received | op_inbytes | Number of bytes received for the SMB write requests. The byte count also includes protocol headers. | bytes/s | ||
Write - Operations | op_count | Number of times the corresponding SMB write request is called. | ops/s | ||
Write - Latency | op_time / op_count | Total amount of time spent for all types of SMB write requests. | s (ns, ms, us) | ||
Negprot - Sent | op_outbytes | Number of bytes sent for the SMB negprot requests. The byte count also includes protocol headers. | bytes/s | ||
Negprot - Received | op_inbytes | Number of bytes received for the SMB negprot requests. The byte count also includes protocol headers. | bytes/s | ||
Negprot - Operations | op_count | Number of times the corresponding SMB negprot request is called. | ops/s | ||
Negprot - Latency | op_time / op_count | Total amount of time spent for all types of SMB negprot requests. | s (ns, ms, us) | ||
Sesssetup – Sent | op_outbytes | Number of bytes sent for the SMB sesssetup requests. The byte count also includes protocol headers. | bytes/s | ||
Sesssetup - Received | op_inbytes | Number of bytes received for the SMB sesssetup requests. The byte count also includes protocol headers. | bytes/s | ||
Sesssetup - Operations | op_count | Number of times the corresponding SMB sesssetup request is called. | ops/s | ||
Sesssetup - Latency | op_time / op_count | Total amount of time spent for all types of SMB sesssetup requests. | s (ns, ms, us) | ||
Logoff - Sent | op_outbytes | Number of bytes sent for the SMB logoff requests. The byte count also includes protocol headers. | bytes/s | ||
Logoff - Received | op_inbytes | Number of bytes received for the SMB logoff requests. The byte count also includes protocol headers. | bytes/s | ||
Logoff - Operations | op_count | Number of times the corresponding SMB logoff request is called. | ops/s | ||
Logoff - Latency | op_time / op_count | Total amount of time spent for all types of SMB logoff requests. | s (ns, ms, us) | ||
Tcon - Sent | op_outbytes | Number of bytes sent for the SMB Tcon requests. The byte count also includes protocol headers. | bytes/s | ||
Tcon - Received | op_inbytes | Number of bytes received for the SMB Tcon requests. The byte count also includes protocol headers. | bytes/s | ||
Tcon - Operations | op_count | Number of times the corresponding SMB Tcon request is called. | ops/s | ||
Tcon - Latency | op_time / op_count | Total amount of time spent for all types of SMB Tcon requests. | ops/s | ||
Tdis - Sent | op_outbytes | Number of bytes sent for the SMB Tdis requests. The byte count also includes protocol headers. | bytes/s | ||
Tdis - Received | op_inbytes | Number of bytes received for the SMB Tdis requests. The byte count also includes protocol headers. | bytes/s | ||
Tdis - Operations | op_count | Number of times the corresponding SMB Tdis request is called. | ops/s | ||
Tdis - Latency | op_time / op_count | Total amount of time spent for all types of SMB Tdis requests. | s (ns, ms, us) | ||
Lock - Sent | op_outbytes | Number of bytes sent for the SMB lock requests. The byte count also includes protocol headers. | bytes/s | ||
Lock - Received | op_inbytes | Number of bytes received for the SMB lock requests. The byte count also includes protocol headers. | bytes/s | ||
Lock - Operations | op_count | Number of times the corresponding SMB lock request is called. | ops/s | ||
Lock - Latency | op_time | Total amount of time spent for all types of SMB lock requests. | bytes/s | ||
IOCtl - Sent | op_outbytes | Number of bytes sent for the SMB IOCtl requests. The byte count also includes protocol headers. | bytes/s | ||
IOCtl - Received | op_inbytes | Number of bytes received for the SMB IOCtl requests. The byte count also includes protocol headers. | bytes/s | ||
IOCtl - Operations | op_count | Number of times the corresponding SMB IOCtl request is called. | ops/s | ||
IOCtl - Latency | op_time / op_count | Total amount of time spent for all types of SMB IOCtl requests. | s (ns, ms, us) | ||
Cancel - Sent | op_outbytes | Number of bytes sent for the SMB cancel requests. The byte count also includes protocol headers. | bytes/s | ||
Cancel - Received | op_inbytes | Number of bytes received for the SMB cancel requests. The byte count also includes protocol headers. | bytes/s | ||
Cancel - Operations | op_count | Number of times the corresponding SMB cancel request is called. | ops/s | ||
Cancel - Latency | op_time / op_count | Total amount of time spent for all types of SMB cancel requests. | s (ns, ms, us) | ||
KeepAlive - Sent | op_outbytes | Number of bytes sent for the SMB KeepAlive requests. The byte count also includes protocol headers. | bytes/s | ||
KeepAlive - Received | op_inbytes | Number of bytes received for the SMB KeepAlive requests. The byte count also includes protocol headers. | bytes/s | ||
KeepAlive - Operations | op_count | Number of times the corresponding SMB KeepAlive request is called. | ops/s | ||
KeepAlive - Latency | op_time / op_count | Total amount of time spent for all types of SMB KeepAlive requests. | s (ns, ms, us) | ||
Find - Sent | op_outbytes | Number of bytes sent for the SMB find requests. The byte count also includes protocol headers. | bytes/s | ||
Find - Received | op_inbytes | Number of bytes received for the SMB find requests. The byte count also includes protocol headers. | bytes/s | ||
Find - Operations | op_count | Number of times the corresponding SMB find request is called. | ops/s | ||
Find - Latency | op_time / op_count | Total amount of time spent for all types of SMB find requests. | s (ns, ms, us) | ||
Notify - Sent | op_outbytes | Number of bytes sent for the SMB notify requests. The byte count also includes protocol headers. | bytes/s | ||
Notify - Received | op_inbytes | Number of bytes received for the SMB notify requests. The byte count also includes protocol headers. | bytes/s | ||
Notify - Operations | op_count | Number of times the corresponding SMB notify request is called. | ops/s | ||
Notify - Latency | op_time / op_count | Total amount of time spent for all types of SMB KeepAlive requests. | s (ns, ms, us) | ||
GetInfo - Sent | op_outbytes | Number of bytes sent for the SMB GetInfo requests. The byte count also includes protocol headers. | bytes/s | ||
GetInfo - Received | op_inbytes | Number of bytes received for the SMB GetInfo requests. The byte count also includes protocol headers. | bytes/s | ||
GetInfo - Operations | op_count | Number of times the corresponding SMB GetInfo request is called. | ops/s | ||
GetInfo - Latency | op_time / op_count | Total amount of time spent for all types of SMB GetInfo requests. | s (ns, ms, us) | ||
SetInfo - Sent | op_outbytes | Number of bytes sent for the SMB SetInfo requests. The byte count also includes protocol headers. | bytes/s | ||
SetInfo - Received | op_inbytes | Number of bytes received for the SMB SetInfo requests. The byte count also includes protocol headers. | bytes/s | ||
SetInfo - Operations | op_count | Number of times the corresponding SMB SetInfo request is called. | ops/s | ||
SetInfo - Latency | op_time / op_count | Total amount of time spent for all types of SMB SetInfo requests. | s (ns, ms, us) | ||
Break - Sent | op_outbytes | Number of bytes sent for the SMB Break requests. The byte count also includes protocol headers. | bytes/s | ||
Break - Received | op_inbytes | Number of bytes received for the SMB Break requests. The byte count also includes protocol headers. | bytes/s | ||
Break - Operations | op_count | Number of times the corresponding SMB Break request is called. | ops/s | ||
Break - Latency | op_time / op_count | Total amount of time spent for all types of SMB Break requests. | s (ns, ms, us) |
Resource type - Object
The following table lists the performance metrics available for the resource type - Object.
Resource type and aggregation levels | Sensor names | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
Object – Node, Cluster | SwiftAccount SwiftContainer SwiftObject SwiftProxy |
Account - Auditor | account_auditor_time | Timing data for individual account database audits. Auditors run in the background on every object node in the cluster and continuously scan the disks to ensure that the data stored on disk is not corrupted. | s (ns, ms, us) |
Account - Reaper | account_reaper_time | Time taken for each reap_account call. When an account is deleted, reapers ensure that the associated objects and containers are deleted from the object storage. | s (ns, ms, us) | ||
Account - DEL | account_DEL_time | Time taken for each successful account DELETE request in a node. | s (ns, ms, us) | ||
Account - PUT | account_PUT_time | Time taken for each successful PUT request at the account level. | s (ns, ms, us) | ||
Account - HEAD | account_HEAD_time | Time taken for each successful HEAD request at the account level. | s (ns, ms, us) | ||
Account - GET | account_GET_time | Time taken for each successful GET request at the account level. | s (ns, ms, us) | ||
Account - REPLICATE | account_REPLICATE_time | Time taken for each successful REPLICATE request at the account level. | s (ns, ms, us) | ||
Account - POST | account_POST_time | Time taken for each successful POST request at the account level. | s (ns, ms, us) | ||
Account - Replicator | account_replicator_time | Time taken for each successful database replication attempt at the account level. | s (ns, ms, us) | ||
Container - Auditor | container_auditor_time | Time taken for individual container audits. Auditors run in the background on every object node in the cluster and continuously scan the disks to ensure that the data stored on disk is not corrupted. | s (ns, ms, us) | ||
Container - Replicator | container_replicator_time | Time taken for each successful database replication attempt at the container level. | s (ns, ms, us) | ||
Container - DEL | container_DEL_time | Time taken for each successful account DELETE request at the container level in a node. | s (ns, ms, us) | ||
Container - PUT | container_PUT_time | Time taken for each successful PUT request at the container level. | s (ns, ms, us) | ||
Container - HEAD | container_HEAD_time | Time taken for each successful HEAD request at the container level. | s (ns, ms, us) | ||
Container - GET | container_GET_time | Time taken for each successful GET request at the container level. | s (ns, ms, us) | ||
Container - REPLICATE | container_REPLICATE_time | Time taken for each successful REPLICATE request at the container level. | s (ns, ms, us) | ||
Container - POST | container_POST_time | Time taken for each successful POST request at the container level. | s (ns, ms, us) | ||
Container - Sync deletes | container_sync_deletes_time | Time taken for each container database row synchronization due to deletion. | s (ns, ms, us) | ||
Container - Sync puts | container_sync_puts_time | Time taken for each container database row synchronization due to the PUT operation. | s (ns, ms, us) | ||
Container - Updater | container_updater_time | Timing data for processing a container; only includes timing for containers which needed to update their accounts. | s (ns, ms, us) | ||
Object - Auditor | object_auditor_time | Timing data for each object audit. This includes rate-limiting sleep time for max_bytes_per_second and it does not include any rate-limiting sleep time for max_files_per_second. | s (ns, ms, us) | ||
Object - Expirer | object_expirer_time | Timing data for each object expiration attempt, including ones resulting in an error. | s (ns, ms, us) | ||
Object - Replicator partition delete | object_replicator_partition_delete_time | Timing data for partitions replicated to another node because they didn’t belong on this node. This metric is not tracked per device. | s (ns, ms, us) | ||
Object - Replicator partition update | object_replicator_partition_update_time | Timing data for partitions replicated which also belong on this node. This metric is not tracked per-device. | s (ns, ms, us) | ||
Object - POST | object_POST_time | Time taken for each successful object POST request. | s (ns, ms, us) | ||
Object - PUT | object_PUT_time | Time taken for each successful object PUT request. | s (ns, ms, us) | ||
Object - GET | object_GET_time | Time taken for each successful object GET request. This includes requests that could not find the object; for example, including disk errors resulting in file quarantine. | s (ns, ms, us) | ||
Object - HEAD | object_HEAD_time | Time taken for each successful object HEAD request. This includes requests that could not find the object; for example, including disk errors resulting in file quarantine. | s (ns, ms, us) | ||
Object - DEL | object_DEL_time | Time taken for each successful object DELETE request. | s (ns, ms, us) | ||
Object - REPLICATE | object_REPLICATE_time | Time taken for each successful object REPLICATE request. | s (ns, ms, us) | ||
Object - Updater | object_updater_time | Time taken for each object sweeps to flush async_pending container updates. This timing data does not include object sweeps which did not find an existing async_pending storage directory. | s (ns, ms, us) | ||
Proxy - Account latency | proxy_account_latency | Time taken to complete sending the response headers at the account level. The standard response code for successful HTTP requests is 200. | s (ns, ms, us) | ||
Proxy - Container latency | proxy_container_latency | Time taken to complete sending the response headers at the container level. The standard response code for successful HTTP requests is 200. | s (ns, ms, us) | ||
Proxy - Object latency | proxy_object_latency | Time taken to complete sending the response headers at the object level. The standard response code for successful HTTP requests is 200. | s (ns, ms, us) | ||
Proxy - Account GET latency | proxy_account_GET_time | Time taken to complete GET requests at the account level. The standard response code for successful HTTP requests is 200. | s (ns, ms, us) | ||
Proxy - Account GET throughput | proxy_account_GET_bytes | Number of bytes transferred in GET requests at the account level. The standard response code for successful HTTP requests is 200. | bytes/s | ||
Proxy - Container GET latency | proxy_container_GET_time | Time taken to complete GET requests at the container level. The standard response code for successful HTTP requests is 200. | s (ns, ms, us) | ||
Proxy - Container GET throughput | proxy_container_GET_bytes | Number of bytes transferred in GET requests at the container level. The standard response code for successful HTTP requests is 200. | bytes/s | ||
Proxy - Container PUT latency | proxy_container_PUT_time | Time taken for each successful PUT request at the container level. The response code '201' indicates that the request has been fulfilled and new resource is created. | s (ns, ms, us) | ||
Proxy - Container PUT throughput | proxy_container_PUT_bytes | Number of bytes transferred in each successful PUT request at the container level. The response code '201' indicates that the request has been fulfilled and new resource is created. | bytes/s | ||
Proxy - Container DEL latency | proxy_container_DEL_time | Time taken to complete each DELETE request at the container level. The response code '204' indicates that the request has been processed and no content is returned. | s (ns, ms, us) | ||
Proxy - Container DEL throughput | proxy_container_DEL_bytes | Number of bytes transferred for all DELETE requests at the container level. The response code '204' indicates that the request has been processed and no content is returned. | bytes/s | ||
Proxy - Object GET latency | proxy_object_GET_time | Time taken to complete GET requests at the object level. The standard response code for successful HTTP requests is 200. | s (ns, ms, us) | ||
Proxy - Object GET throughput | proxy_object_GET_bytes | Number of bytes transferred in GET requests at the object level. The standard response code for successful HTTP requests is 200. | bytes/s | ||
Proxy - Object PUT latency | proxy_object_PUT_time | Time taken for each successful PUT request at the object level. The response code '201' indicates that the request has been fulfilled and new resource is created. | s (ns, ms, us) | ||
Proxy - Object PUT throughput | proxy_object_PUT_bytes | Number of bytes transferred in each successful PUT request at the object level. The response code '201' indicates that the request has been fulfilled and new resource is created. | bytes/s | ||
Proxy - Object DEL latency | proxy_object_DEL_time | Time taken to complete each DELETE request at the object level. The response code '204' indicates that the request has been processed and no content is returned. | s (ns, ms, us) | ||
Proxy - Object DEL throughput | proxy_object_DEL_bytes | Number of bytes transferred for all DELETE requests at the object level. The response code '204' indicates that the request has been processed and no content is returned. | bytes/s |
Resource type - CTDB
The following table lists the performance metrics available for the resource type - CTDB.
Resource type and aggregation levels | Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
CTDB - Node | CTDBStats | CTDB DB Stats - RO delegations | db_ro_delegations | Number of CTDB read-only delegations created. | 1/s |
CTDB DB Stats - RO revokes | db_ro_revokes | Number of CTDB read-only delegations that were revoked. The difference between number of total revokes and delegations gives the number of currently active read-only delegations. | 1/s | ||
CTDB DB Stats - Locks num. calls | db_locks_num_calls | Number of completed lock calls. This includes database locks and record locks. | 1/s | ||
CTDB DB Stats - Locks num. failed | db_locks_num_failed | Number of failed lock calls. This includes database locks and record locks. | 1/s | ||
CTDB DB Stats - Locks num. current | db_locks_num_current | Number of scheduled lock calls. This includes database locks and record locks. | 1/s | ||
CTDB DB Stats - Locks num. pending | db_locks_num_pending | Number of queued lock calls. This includes database locks and record locks. | 1/s | ||
CTDB Stats - Num. clients | num_clients | Number of processes currently connected to CTDB's Unix socket. This includes recovery daemon, CTDB tool and, SMB processes such as smbd and winbindd. | 1/s | ||
CTDB Stats - Num. recoveries | num_recoveries | Number of recoveries since the start of CTDB or since the last statistics reset. | 1/s | ||
CTDB Stats - Client packets sent | client_packets_sent | Number of packets sent to client processes through Unix domain socket. | 1/s | ||
CTDB Stats - Client packets received | client_packets_recv | Number of packets received from client processes through Unix domain socket. | 1/s | ||
CTDB Stats - Node packets sent | node_packets_sent | Number of packets sent to other nodes in the cluster through TCP. | 1/s | ||
CTDB Stats - Node packets received | node_packets_recv | Number of packets received from other nodes in the cluster through TCP. | 1/s | ||
CTDB Stats - Keepalive packets sent | keepalive_packets_sent | Number of keepalive messages sent to other nodes. CTDB periodically sends keepalive messages to other nodes. | 1/s | ||
CTDB Stats - Keeplalive packets received | keepalive_packets_recv | Number of keepalive messages received from other nodes. | 1/s | ||
CTDB Stats - Node req. call | node_req_call | Number of REQ_CALL messages from other nodes. | 1/s | ||
CTDB Stats - Node reply call | node_reply_call | Number of REPLY_CALL messages from the other nodes. | 1/s | ||
CTDB Stats - Node req. dmaster | node_req_dmaster | Number of REQ_DMASTER messages from other nodes. | 1/s | ||
CTDB Stats - Node reply dmaster | node_reply_dmaster | Number of REPLY_DMASTER messages from other nodes. | 1/s | ||
CTDB Stats - Node reply error | node_reply_error | Number of REPLY_ERROR messages from theother nodes. | 1/s | ||
CTDB Stats - Node req. message | node_req_message | Number of REQ_MESSAGE messages from the clients. | 1/s | ||
CTDB Stats - Node req. control | node_req_control | Number of REQ_CONTROL messages from the clients. | 1/s | ||
CTDB Stats - Node reply control | node_reply_control | Number of REPLY_CONTROL messages from other nodes. | 1/s | ||
CTDB Stats - Client req. call | client_req_call | Number of REQ_CALL messages from the clients. | 1/s | ||
CTDB Stats - Client req. message | client_req_message | Number of REQ_MESSAGE messages from the clients. | 1/s | ||
CTDB Stats - Client req. control | client_req_control | Number of timeouts for REQ_CONTROL messages. | 1/s | ||
CTDB Stats - Timeouts call | timeouts_call | Number of timeouts for REQ_CALL messages. | 1/s | ||
CTDB Stats - Timeouts control | timeouts_control | Number of timeouts for REQ_CONTROL messages. | 1/s | ||
CTDB Stats - Timeouts traverse | timeouts_traverse | Number of timeouts for database traverse operations. | 1/s | ||
CTDB Stats - Locks num. calls | locks_num_calls | Number of completed lock calls. This includes database locks and record locks. | 1/s | ||
CTDB Stats - Locks num. current | locks_num_current | Number of scheduled lock calls. This includes database locks and record locks. | 1/s | ||
CTDB Stats - Locks num. pending | locks_num_pending | Number of queued lock calls. This includes database locks and record locks. | 1/s | ||
CTDB Stats - Locks num. failed | locks_num_failed | Number of failed lock calls. This includes database locks and record locks. | 1/s | ||
CTDB Stats - Total calls | total_calls | Number of REQ_CALL messages processed from clients. This number should be the same as client's REQ_CALL | 1/s | ||
CTDB Stats - Pending calls | pending_calls | Number of REQ_CALL messages that are currently being processed. This number indicates the number of record migrations in flight. | 1/s | ||
CTDB Stats - Childwrite calls | childwrite_calls | Number of record update calls. Record update calls are used to update a record that is under transaction. | 1/s | ||
CTDB Stats - Pending childwrite calls | pending_childwrite_calls | Number of record update calls that are currently active. | 1/s | ||
CTDB Stats - Memory used | memory_used | The amount of memory in bytes currently used by CTDB using talloc. This includes all the memory used for CTDB´s internal data structures. This does not include the memory mapped TDB databases. | MiB | ||
CTDB Stats - Max hop count | max_hop_count | The maximum number of hops required for a record migration request to obtain the record. High numbers indicate the record contention. | Number; so no units. | ||
CTDB Stats - Total RO delegations | total_ro_delegations | Number of read-only delegations created. | Number; so no units. | ||
CTDB Stats - Total RO revokes | total_ro_revokes | Number of read-only delegations that were revoked. The difference between total_ro_revokes and total_ro_delegations gives the number of currently active read-only delegations. | Number; so no units. | ||
CTDB Stats - Locks latency min. | locks_latency_min | Minimum time that is required to obtain the record locks. | s (ns, ms, us) | ||
CTDB Stats - Locks latency max. | locks_latency_max | Maximum time that is required to obtain the record locks. | s (ns, ms, us) | ||
CTDB Stats - Locks latency avg. | locks_latency_avg | Average time that is required to obtain the record locks. | s (ns, ms, us) | ||
CTDB Stats - Locks latency num. | locks_latency_num | Number of record lock requests. | |||
CTDB Stats - Call latency min. | call_latency_min | Minimum time that is required to process a REQ_CALL message from client. This includes the time required to migrate a record from remote node, if the record is not available on the local node. | s (ns, ms, us) | ||
CTDB Stats - Call latency max. | call_latency_max | Maximum time that is required to process a REQ_CALL message from client. This includes the time required to migrate a record from remote node, if the record is not available on the local node. | s (ns, ms, us) | ||
CTDB Stats - Call latency avg. | call_latency_avg | The average time that is required to process a REQ_CALL message from client. This includes the time required to migrate a record from remote node, if the record is not available on the local node. | s (ns, ms, us) | ||
CTDB Stats - Call latency num. | call_latency_num | Number of REQ_CALL messages from the client. | |||
CTDB Stats - Childwrite latency min. | childwrite_latency_min | Minimum time that is required to update records under a transaction. | s (ns, ms, us) | ||
CTDB Stats - Childwrite latency max. | childwrite_latency_max | Maximum time that is required to update records under a transaction. | s (ns, ms, us) | ||
CTDB Stats - Childwrite latency avg. | childwrite_latency_avg | Average time that is required to update records under a transaction. | s (ns, ms, us) | ||
CTDB Stats - Childwrite latency num. | childwrite_latency_num | Number of record update calls. | |||
CTDB Stats - Reclock ctdbd min. | reclock_ctdbd_min | Minimum time that is required to check if recovery lock is still held by the recovery daemon when recovery mode is changed. This check is done in CTDB daemon. | s (ns, ms, us) | ||
CTDB Stats - Reclock ctdbd max. | reclock_ctdbd_max | The maximum time that is required to check if recovery lock is still held by the recovery daemon when recovery mode is changed. This check is done in CTDB daemon. | s (ns, ms, us) | ||
CTDB Stats - Reclock ctdbd avg. | reclock_ctdbd_avg | The average time that is required to check if recovery lock is still held by the recovery daemon when recovery mode is changed. This check is done in CTDB daemon. | s (ns, ms, us) | ||
CTDB Stats - Reclock ctdbd num. | reclock_ctdbd_num | Number of checks if the recovery lock is still held by the recovery daemon when recovery mode is changed. | Number; so no units. | ||
CTDB - Database per Node, Cluster | CTDB DB Stats - RO delegations | db_ro_delegations | Number of read-only delegations created. | Number; so no units. | |
CTDB DB Stats - RO revokes | db_ro_revokes | Number of read-only delegations that were revoked. The difference between total_ro_revokes and total_ro_delegations gives the number of currently active read-only delegations. | Number; so no units. | ||
CTDB DB Stats - Locks num. calls | db_locks_num_calls | Number of completed lock calls. This includes database locks and record locks. | Number; so no units. | ||
CTDB DB Stats - Locks num. failed | db_locks_num_failed | Number of failed lock calls. This includes database locks and record locks. | Number; so no units. | ||
CTDB DB Stats - Locks num. current | db_locks_num_current | Number of scheduled lock calls. This includes database locks and record locks. | Number; so no units. | ||
CTDB DB Stats - Locks num. pending | db_locks_num_pending | Number of queued lock calls. This includes database locks and record locks. | Number; so no units. |
Resource type - Transparent Cloud Tiering
The following table lists the performance metrics available for the resource type - Transparent Cloud Tiering.
Resource type and aggregation levels | Sensor names | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
Transparent cloud tiering - Node and Cluster |
TCTFsIcstoreStats TCTFsetIcstoreStats TCTFsGpfsConnectorStats TCTFsetGpfsConnectorStats tct_fset_total_failed_requests |
Migrate | MCStoreGPFSStats|migrate|mcs_total_requests | Total number of migrate requests. | ops/s |
MCStoreGPFSStats|migrate|mcs_total_request_time | Total time taken for migrate requests. | ms | |||
MCStoreGPFSStats|migrate|mcs_total_failed_requests | Total number of failed migrate requests. | ops/s | |||
MCStoreGPFSStats|migrate|mcs_total_failed_requests_time | Total time spent in failed migrate requests. | ms | |||
MCStoreIcstoreStats|migrate|mcs_total_bytes | Total number of bytes transferred from file system to the cloud provider. | bytes/s | |||
MCStoreIcstoreStats|migrate|mcs_total_persisted_bytes | The total number of transferred bytes that are successfully persisted on the cloud provider. | bytes/s | |||
MCStoreIcstoreStats|migrate|mcs_total_retried_operations | Total number of retry PUT operations. | ops/s | |||
MCStoreIcstoreStats|migrate|mcs_total_operation_errors | Total number of error PUT operations. | ops/s | |||
Recall | MCStoreGPFSStats|recall|mcs_total_requests | Total number of recall requests. | ms | ||
MCStoreGPFSStats|recall|mcs_total_request_time | Total time taken for recall requests. | ops/s | |||
MCStoreGPFSStats|recall|mcs_total_requests | Total number of failed recall requests. | ops/s | |||
MCStoreGPFSStats|recall|mcs_total_failed_requests_time | Total time spent in failed recall requests. | ms | |||
MCStoreIcstoreStats|recall|mcs_total_bytes | Total number of bytes downloaded from the cloud provider to the file system. | bytes/s | |||
MCStoreIcstoreStats|recall|mcs_total_persisted_bytes | Total number of downloaded bytes that are successfully persisted on the file system. | bytes/s | |||
MCStoreIcstoreStats|recall|mcs_total_retried_operations | Total number of retry GET operations. | ops/s | |||
MCStoreIcstoreStats|recall|mcs_total_operation_errors | Total number of error GET operations. | ops/s | |||
Remove | MCStoreGPFSStats|remove|mcs_total_requests | Total number of remove requests. | ops/s | ||
MCStoreGPFSStats|remove|mcs_total_request_time | Total time taken for remove requests. | ops/s | |||
MCStoreGPFSStats|remove|mcs_total_requests | Total number of failed remove requests. | ops/s | |||
MCStoreGPFSStats|remove|mcs_total_failed_requests_time | Total time spent in failed remove requests. | ms |
Resource type - Waiters
The following table lists the performance metrics available for the resource type - Waiters.
Resource type and aggregation levels | Sensor names | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
Waiters - Cluster | GPFSWaiters | All | gpfs_wt_count_all | All waiter types | s (ns, ms, us) |
Network I/O (RPC waiters) | gpfs_wt_count_network_io | All waiters that concern network I/O operations of type RPC calls. | s (ns, ms, us) | ||
Local disk I/O waiters | gpfs_wt_count_local_io | All waiters that concern local disk I/O operations. | s (ns, ms, us) | ||
Waiters (ThMutex) | gpfs_wt_count_thmutex | Waiters of the type ThMutex | s (ns, ms, us) | ||
Waiters (Syscall) | gpfs_wt_count_syscall | Waiters of the type Syscall | s (ns, ms, us) | ||
Waiters (ThCond) | gpfs_wt_count_thcond | Waiters of the type ThCond | s (ns, ms, us) | ||
Waiters (Delay) | gpfs_wt_count_delay | Waiters of the type Delay | s (ns, ms, us) |
Resource type - AFM
The following table lists the performance metrics available for the resource type - AFM
Resource type and aggregation levels | Sensor names | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|---|
Node, Cluster | GPFSAFM | Bytes Read | gpfs_afm_fset_bytes_read | Total number of bytes read from home to cache as a result of cache miss. | bytes/s |
Bytes Written | gpfs_afm_bytes_written | Total number of bytes written to the home as a result of cache updates. | bytes/s | ||
Operations – Total | gpfs_afm_ops_sent | Total number of operations sent over the communication protocol to the remote system. | ops/s | ||
Operations – Synchronous | gpfs_afm_ops_sync | Number of synchronous operations that are sent to the remote system. | ops/s | ||
Operations – Expired | gpfs_afm_ops_expired | Number of operations sent to remote system as they are expired. One operation is marked as expired when the configured async timeout is exceeded in the gateway queue. | ops/s | ||
Operations – Forced | gpfs_afm_ops_forced | Number of operations sent to remote system as they are forced out of the gateway queue before the configured async timeout. This might be because of a dependent operation. | ops/s | ||
Operations – Revoke | gpfs_afm_ops_revoked | Number of operations sent to the remote system because a conflicting token acquired from another IBM Spectrum Scale node resulted in a revoke. | ops/s | ||
Queue – Backlog | gpfs_afm_bytes_pending | Total number of bytes that are not yet written to the remote system. | bytes | ||
Queue – Messages | gpfs_afm_num_queued_msgs | Number of messages that are currently enqueued. | number, no unit | ||
Queue – Memory | gpfs_afm_used_q_memory | Memory used in bytes for the enqueued messages. | bytes | ||
Latency -Min Queue Wait | gpfs_afm_shortest_time | Shortest time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
Latency – Avg Queue Wait | gpfs_afm_avg_time | Average time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
Latency – Max Queue Wait | gpfs_afm_longest_time | Longest time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
File System | GPFSAFMFS | Bytes Read | gpfs_afm_fs_bytes_read | Total number of bytes read from remote system as a result of cache miss. | bytes/s |
Bytes Written | gpfs_afm_fs_bytes_written | Total number of bytes written to the home as a result of cache updates. | bytes/s | ||
Operations – Total | gpfs_afm_fs_ops_sent | Total number of operations sent over the communication protocol to the remote system. | ops/s | ||
Operations – Synchronous | gpfs_afm_fs_ops_sync | Number of synchronous operations that are sent to the remote system. | ops/s | ||
Operations – Expired | gpfs_afm_fs_ops_expired | Number of operations sent to remote system as they are expired. One operation is marked as expired when the configured async timeout is exceeded in the gateway queue. | ops/s | ||
Operations – Forced | gpfs_afm_fs_ops_forced | Number of operations sent to remote system as they are forced out of the gateway queue before the configured async timeout. This might be because of a dependent operation. | ops/s | ||
Operations – Revoke | gpfs_afm_fs_ops_revoked | Number of operations sent to the remote system because a conflicting token acquired from another IBM Spectrum Scale node resulted in a revoke. | ops/s | ||
Queue – Backlog | gpfs_afm_fs_bytes_pending | Total number of bytes that are not yet written to the remote system. | bytes | ||
Queue – Messages | gpfs_afm_fs_num_queued_msgs | Number of messages that are currently enqueued. | number, no unit | ||
Queue – Memory | gpfs_afm_fs_used_q_memory | Used memory in bytes by the messages enqueued. | bytes | ||
Latency -Min Queue Wait | gpfs_afm_fs_shortest_time | Shortest time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
Latency – Avg Queue Wait | gpfs_afm_fs_avg_time | Average time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
Latency – Max Queue Wait | gpfs_afm_fs_longest_time | Longest time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
Fileset | GPFSAFMFSET | Bytes Read | gpfs_afm_fset_bytes_read | Total number of bytes read from home to cache as a result of cache miss. | bytes/s |
Bytes Written | gpfs_afm_fset_bytes_written | Total number of bytes written to the home as a result of cache updates. | bytes/s | ||
Operations – Total | gpfs_afm_fset_ops_sent | Total number of operations sent over the communication protocol to the remote system. | ops/s | ||
Operations – Synchronous | gpfs_afm_fset_ops_sync | Number of synchronous operations that are sent to the remote system. | ops/s | ||
Operations – Expired | gpfs_afm_fset_ops_expired | Number of operations sent to remote system as they are expired. One operation is marked as expired when the configured async timeout is exceeded in the gateway queue. | ops/s | ||
Operations – Forced | gpfs_afm_fset_ops_forced | Number of operations sent to remote system as they are forced out of the gateway queue before the configured async timeout. This might be because of a dependent operation. | ops/s | ||
Operations – Revoke | gpfs_afm_fset_ops_revoked | Number of operations sent to the remote system because a conflicting token acquired from another IBM Spectrum Scale node resulted in a revoke. | ops/s | ||
Queue – Backlog | gpfs_afm_fset_bytes_pending | Total number of bytes that are not yet written to the remote system. | bytes | ||
Queue – Messages | gpfs_afm_fset_num_queued_msgs | Number of messages that are currently enqueued. | number, no unit | ||
Queue – Memory | gpfs_afm_fset_used_q_memory | Used memory in bytes by the messages enqueued. | bytes | ||
Latency -Min Queue Wait | gpfs_afm_fset_shortest_time | Shortest time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
Latency – Avg Queue Wait | gpfs_afm_fset_avg_time | Average time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) | ||
Latency – Max Queue Wait | gpfs_afm_fset_longest_time | Longest time in seconds that a pending operation waited in the gateway queue before being sent to the remote system. | s (ms, s) |
Capacity metrics
The following table lists the performance metrics available for analyzing the capacity details at various levels.
Sensor name | Metric name in GUI | Corresponding mmperfmon metric name | Description | Unit |
---|---|---|---|---|
GPFSCapacity | Total capacity | gpfs_disk_disksize | Overall capacity of the NSD or file system. | Bytes (KiB, MiB, GiB, TiB) |
Available capacity in fragments | gpfs_disk_free_fragkb | Available in fragments. | Bytes (KiB, MiB, GiB, TiB) | |
Available capacity in full blocks | gpfs_disk_free_fullkb | Available capacity in full blocks. | Bytes (KiB, MiB, GiB, TiB) | |
Total | gpfs_pool_disksize | Overall capacity of the pool. | Bytes (KiB, MiB, GiB, TiB) | |
Available capacity in full blocks | gpfs_pool_free_fullkb | Available capacity of the pool in full blocks. | Bytes (KiB, MiB, GiB, TiB) | |
Allocated inodes | gpfs_fs_inode_alloc | Number of allocated inodes in the file system that are aggregated over all inode spaces. | Bytes (KiB, MiB, GiB, TiB) | |
Available inodes | gpfs_fs_inode_free | Number of available inodes in the file system that are aggregated over all inode spaces. | Bytes (KiB, MiB, GiB, TiB) | |
Max inodes | gpfs_fs_inode_max | Maximum number of inodes in the file system that are aggregated over all inode spaces. | Bytes (KiB, MiB, GiB, TiB) | |
DiskFree | Total capacity | df_total | Total capacity of a local file system. | Bytes (KiB, MiB, GiB, TiB) |
Available capacity | df_free | Available capacity in a local file system. | Bytes (KiB, MiB, GiB, TiB) | |
Used capacity | df_used | Used capacity in a local file system. | Bytes (KiB, MiB, GiB, TiB) | |
GPFSFilesetQuota | Capacity | gpfs_rq_blk_current | Used capacity in a fileset, as reported by the quota mechanism. | Bytes (KiB, MiB, GiB, TiB) |
Capacity hard limit | gpfs_rq_blk_hard_limit | Hard quota limit for capacity in a fileset. When this limit is exceeded, adding more data to the fileset will be prevented. | Bytes (KiB, MiB, GiB, TiB) | |
Capacity soft limit | gpfs_rq_blk_soft_limit | Soft quota limit for capacity in a fileset. When this limit is exceeded, adding more data to the fileset will be prevented after a grace period. | Bytes (KiB, MiB, GiB, TiB) | |
Inodes | gpfs_rq_file_current | Used inodes in a fileset, as reported by the quota mechanism. | Bytes (KiB, MiB, GiB, TiB) | |
Inodes hard limit | gpfs_rq_file_hard_limit | Hard quota limit for number of inodes in a fileset. When this limit is exceeded, adding more files, links, or directories to the fileset will be prevented. | Bytes (KiB, MiB, GiB, TiB) | |
Inodes soft limit | gpfs_rq_file_soft_limit | Soft quota limit for number of inodes in a fileset. When this limit is exceeded, adding more files, links, or directories to the fileset will be prevented after a grace period. | Bytes (KiB, MiB, GiB, TiB) |