Creating a IBM Spectrum Scale Client component with the IBM Spectrum Scale Client Policy

Use the Pattern Builder to create virtual application patterns or virtual system patterns with the IBM Spectrum Scale Client Policy for deploying a IBM Spectrum Scale Client component.

1. Create a virtual application pattern or virtual system pattern by using the Pattern Builder.
   Configure your pattern as needed by providing the necessary information in the pattern properties pane and save your changes.
2. For each pattern component that is being configured as a IBM Spectrum Scale Client, select the GPFS Client Policy from the Add Policies list to add it to the pattern component.
3. Select each IBM Spectrum Scale Client pattern component in the Pattern Builder canvas, and configure the following settings in the properties pane:

   IBM Spectrum Scale Client Configurations section

   You can select from one of the following options for the IBM Spectrum Scale Client deployment:

   • IBM Spectrum Scale Client Install and Configuration

     Select this option to install the IBM Spectrum Scale software on the deployed component instance virtual machine, and configure it for use with a IBM Spectrum Scale server that was previously deployed from one of the provided IBM Spectrum Scale Server patterns. If you select this option, you are prompted to provide additional configuration information for one or more file system mounts.

   • IBM Spectrum Scale Client Install for off rack IBM Spectrum Scale Server

     This option installs only the IBM Spectrum Scale software on the deployed component instance virtual machine. It does not attempt to connect to a IBM Spectrum Scale server to set up the access to the shared file system.

     You can use this option if you want to connect to an external IBM Spectrum Scale server outside the Cloud Pak System environment or if you want to connect to a IBM Spectrum Scale server deployed on Cloud Pak System, but after the client instance is deployed. For more information, see the Related links.

   IBM Spectrum Scale Manager IP and Client Key

   Specify the IBM Spectrum Scale Manager IP address and Client Key if you are not using the IBM Spectrum Scale shared service.

   Client Cluster options

   Client maxblocksize

   This field configures the client cluster maxblocksize configuration parameter. If the client cluster maxblocksize configuration parameter is smaller than the IBM Spectrum Scale shared file system block size, the client cannot mount the file system. Update this value only when the client fails to mount the file system because the client maxblocksize is smaller than the file system block size. In this case, set the client maxblocksize to the same value as file system block size.

   For more information about the maxblocksize property, see File System Planning and Block size.

   Enforce file set Quota On Root

   This field configures the client cluster enforceFilesetQuotaOnRoot configuration parameter. Select this property (set it to true) if you want to enforce the file set quota on the root user, for this client cluster only. Do not select the property (set it to false) if you do not want to enforce this quota on the root user. The default value is false (check box is not selected). This property applies only to the current client instance and affects only data that is written by the root user from this client cluster. If you want to enforce the file set quota on root for other client instances, you need to set this property on these clients.

   Use this option with caution. If the quota is enforced on the root user, any processes that are executed by the root user on this client fail if the hard quota is exceeded on a file set during the process execution. For more on IBM Spectrum Scale quotas, see Managing GPFS quotas.

   File system mount section(s)

   Four identical configuration sections are provided for you to specify up to four file sets for use during deployment. You can optionally specify additional file sets if needed by using the Additional file system mounts section.

   After deployment, you can also specify additional file sets by using the Allocate directories on shared file server operation (see the Manage > Operations > GPFS Clients operations view in the Instance Console.

   You can select any number of the File system mount sections and provide the required information. If you select multiple sections, ensure that you select them in order from top to bottom.

   You are prompted for the following file system mount properties:

   IBM Spectrum Scale File System Information

   The name of a file system that is defined on the IBM Spectrum Scale Server.

   The IBM Spectrum Scale Server Get Cluster Status operation provides a list of defined file systems and associated file sets. You can define a new file system by using the IBM Spectrum Scale Server Attach Storage Volumes operation.

   IBM Spectrum Scale Fileset directory

   The name of the IBM Spectrum Scale file set directory. A file set is created within the specified file system if the file set does not already exist.

   Note: Blank spaces are not allowed in the file set name, and avoid using the text string IBM Spectrum Scale in the name.

   Storage Maximum limit (K,M,G)

   The maximum amount of data that the file set can use. The storage maximum limit represents the IBM Spectrum Scale hard quota for the file set.

   Data will not be able to be written to the file set when the maximum limit (hard quota) threshold is met. For more information about IBM Spectrum Scale quotas, see Managing GPFS quotas.

   Specify the unit in kilobytes, megabytes, or gigabytes by adding K, M, or G as a suffix. For example, specify 100M for 100 megabytes. Specify the size as a positive integer with no preceding plus (+) character. Enter a 0 value for no maximum limit. Example of valid entries: 20M, 1024K, 1G, 0M.

   When a new file set is created and the storage limit is set, the following command is executed. This command sets no minimum limit (or soft quota) for the specified file set. It sets a maximum limit (hard quota) for the file set.

   su - gpfsprod -c 'sudo /usr/lpp/mmfs/bin/mmsetquota -j <Fileset  Name> -s 0M -h <Maximum Quota Size> <File System Name>

   For more information, see mmsetquota command.

   Note:

   • By default, the file set minimum and maximum limits (or soft and hard quota) are not enforced for the root user. If you want to have the root user follow the specified limits, you need to update the IBM Spectrum Scale cluster to allow that. Run this command on the IBM Spectrum Scale client cluster where you want to enforce the quota for the root user:

     su - gpfsprod -c '/usr/lpp/mmfs/bin/mmchconfig enforceFilesetQuotaOnRoot=yes -i’

     This property applies only to the IBM Spectrum Scale cluster where the command is executed and affects only data that is written by the root user from this cluster. Each IBM Spectrum Scale client instance configures a separate IBM Spectrum Scale client cluster for each virtual machine. If you want to enforce the file set quota on root for other IBM Spectrum Scale client instances, you need to run this command on each of the IBM Spectrum Scale client clusters. For more information, see mmchconfig command.
   • Before you change the IBM Spectrum Scale server cluster to enforce the file set quotas on the root user, make sure that this change does not affect any of the cluster scenarios. Keep in mind that if the quota is enforced for the root user, any process that is executed as a root user that would temporarily write data over the allowed threshold will now fail.

   You must manage the storage quota to ensure that the specified storage limit does not exceed the size of the file system.

   Directory to link on local system

   A symbolic link to the file set is created on the virtual machine file system at the location specified. The value specified must be a syntactically correct directory name, starting with a forward slash (/) character (Example: /myFileSetLink. Spaces are not allowed in the link name.

   Additional file system mounts section

   Use this section to define additional file sets to include in the deployment. You must provide a syntactically correct JSON array in the space provided.

   The following example defines two file systems, fsys_a1 and fsys_b1:

   • File system fsys_a1 contains two file sets, fset_a1 and fset_a2.
     • File set fset_a1 has a quota of 200 megabytes, and a symbolic link is created on the virtual machine at directory location /fset_a1Link.
     • File set fset_a21 has a quota of 201 megabytes, and a symbolic link is created on the virtual machine at directory location /fset_a2Link.
   • File system fsys_b1 contains one file set, fset_b1.
     • File set fset_b1 has a quota of 200 megabytes, and a symbolic link is created on the virtual machine at directory location /fset_b1Link.

   The corresponding JSON array is specified as follows (note that blank spaces are not allowed in the name, maxsize, or link fields):

   [{
     'fileSystem' : 'fsys_a1',
     'fileSet' : [
       {
         'name' : 'fset_a1',
         'maxSize' : '200M',
         'link' : '/fset_a1Link'
       },
       {
         'name' : 'fset_a2',
         'maxSize' : '201M,'
         'link' : '/fset_a2Link'
       }
      ]
     },
    {
     'fileSystem' : 'fsys_b1',
     'fileSet' : [
       {
         'name' : 'fset_b1',
         'maxSize' : '200M',
         'link' : '/fset_b1Link'
       }
      ]
     }]

4. Click Save.