HSCL03F5 HSCE2210 and HSCL0637 errors due to pending and current values out of sync

Resolving The Problem

Problem Description
Activating a partition fails with the following message: HSCL03F5 - There is an insufficient amount of shared processing resources. The error occurs even though the system properties show that Available processing units is greater than the minimum processing units for the profile. For example, Available processing units is .2, and the profiles minimum is .1; however, the error states that there are not .1 processor units available.

Dynamic logical partitioning (dlpar) may fail with HSCE2210 "DLPAR Add processor resources to partition failed on managed system." and HSCL0637 "Unable to modify processing information of the partition. The information could be out of sync. Rebuild the managed system and try again".

Problem Detail
The HMC may return the HSCL03F5 error because it believes there is already a pending request for the available processor that will decrease the current available below the needed minimum value. Likewise, dlpar may fail with errors HSCE2210 and HSCL0637 when pending and current values are not in sync.

The HMC GUI does not show information on pending requests, only current. To check for any pending requests, use the command lshwres and compare the current and pending available values.

To see the system level values run:

lshwres -r proc -m managed-system --level sys

To see the partition level values run:
lshwres -r proc -m managed-system --level lpar

For further information on use of the commands see the "man pages" by running the command man lshwres.  

In the system level example below, the current value is greater than the pending value. The pending value indicates that the pending processor move would leave 0 processor units available:

configurable_sys_proc_units=2.0,curr_avail_sys_proc_units=0.2,pend_avail_sys_proc_units=0.0,installed_sys_proc_units=4.0,deconfig_sys_proc_units=0.0,min_proc_units_per_virtual_proc=0.1,max_virtual_procs_per_lpar=64,max_shared_proc_pools=1

A mismatch between the pending and current values usually occurs when the HMC is unable to restore the settings after a previous dynamic LPAR request failed. This may occur due to the HMC losing the connection to the managed system or partition during the request or due to an error on the server.

Problem Resolution
To force a re-synchronization of the pending and current available values, run the command rsthwres as follows:

rsthwres -r proc -m managed-system

This will re-synchronize all pending values to the current value.