Directed maintenance procedures for tip events

The directed maintenance procedures (DMPs) assist you to repair a problem when you select the action Run fix procedure on a selected event from the GUI > Monitoring > Tips page. DMPs are present for the following tip events reported in the system.

Attention:

If you run these DMPs manually on the command line, the tip event will not reset immediately.

Table 1. Tip events list
Reporting component Event Name Prerequisites Conditions Fix Procedure

gpfs

gpfs_pagepool_ small

gpfs_pagepool_ ok

The actively used GPFS pagepool setting (mmdiag --config | grep pagepool) is lower than or equal to 1 GB.
  • To change the value and make it effective immediately, use the following command:: 
    mmchconfig pagepool=<value> -i

    where <value> is a value higher than 1GB.

  • To change the value and make it effective after next GPFS recycle, use the following command::
    mmchconfig pagepool=<value>

    where <value> is a value higher than 1GB.

  • To ignore the event, use the following command:
    mmhealth event hide gpfs_pagepool_small
AFM component

afm_sensors_ inactive

afm_sensors_ active

 Verify that the node has a gateway designation and a perfmon designation using the mmlscluster command. The period for at least one of the following AFM sensors' is set to 0: GPFSAFM, GPFSAFMFS, GPFSAFMFSET.
  • To change the period when the sensors are defined in the perfmon configuration file, use the following command: 
    mmperfmon config update <sensor_name>.period=<interval>

    Where <sensor_name> is one of the AFM sensors GPFSAFM, GPFSAFMFS, or GPFSAFMFSET, and <interval> is the time in seconds that the sensor waits to gather the different sensors' metrics again.

  • To change the period when the sensors are not defined in the perfmon configuration file, create a sensorsFile with input using the following command: 
    
sensors = {

name = <sensor_name>

period = <interval>

type = "Generic"

}
mmperfmon config add --sensors <path_to_tmp_cfg_file>
  • To ignore the event, use the following command:
    mmhealth event hide afm_sensors_inactive
NFS component

nfs_sensors_ inactive

nfs_sensors_ active

 Verify that the node is NFS enabled, and has a perfmon designation using the mmlscluster command. The NFS sensor NFSIO has a period of 0.
  • To change the period when the sensors are defined in the perfmon configuration file, use the following command: 
    mmperfmon config update <sensor_name>.period=<interval>

    Where <sensor_name> is the NFS sensor NFSIO, and <interval> is the time in seconds that the sensor waits to gather the different sensors' metrics again.

  • To change the period when the sensors are not defined in the perfmon configuration file, use the following command: 
    
mmperfmon config add --sensors /opt/IBM/zimon/defaults/GaneshaProxy.conf
  • To ignore the event, use the following command:
    mmhealth event hide nfs_sensors_inactive
SMB component

smb_sensors_ inactive

smb_sensors_ active

 Verify that the node is SMB enabled, and has a perfmon designation using the mmlscluster command. The period of at least one of the following SMB sensors' is set to 0: SMBStats, SMBGlobalStats .
  • To change the period when the sensors are defined in the perfmon configuration file, use the following command: 
    mmperfmon config update <sensor_name>.period=<interval>

    Where <sensor_name> is one of the SMB sensors SMBStats or SMBGlobalStats, and <interval> is the time in seconds that the sensor waits to gather the different sensors' metrics again.

  • To change the period when the sensors are not defined in the perfmon configuration file, use the following command: 
    
mmperfmon config add --sensors /opt/IBM/zimon/defaults/ZIMonSensors_smb.cfg
  • To ignore the event, use the following command:
    mmhealth event hide smb_sensors_inactive
gpfs

gpfs_maxfilestocache_ small

gpfs_maxfilestocache _ok

 Verify that the node is in the cesNodes node class using the 
mmlsnodeclass --all
command.		 The actively used GPFS maxFilesToCache (mmdiag --config | grep maxFilesToCache) setting has a value smaller than or equal to 100,000.
  • To change the value, use the following command: 
    mmchconfig maxFilesToCache=<value>; mmshutdown; mmstartup

    where <value> is a value higher than 100,000

  • To ignore the event, use the following command: 
    mmhealth event hide gpfs_maxfilestocache_small
gpfs

gpfs_maxstatcache _high

gpfs_maxstatcache _ok

 Verify that the node is a Linux® node. The actively used GPFS maxStatCache (mmdiag --config | grep maxStatCache) value is higher than 0.
  • To change the value, use the following command: 
    mmchconfig maxStatCache=0; mmshutdown; mmstartup
  • To ignore the event, use the following command: 
    mmhealth event hide gpfs_maxstatcache_high
gpfs

callhome_not_enabled

callhome_enabled

 Verify that the node is the Cluster Manager using the mmlsmgr -c command. Call home is not enabled on the cluster.
  • To install call home, install the gpfs.callhome-ecc-client-{version-number}.noarch.rpm package for the ECCClient on the potential call home nodes.
  • To configure the call home package that are installed but not configured:
    1. Issue the mmcallhome capability enable command to initialize the configuration.
    2. Issue the mmcallhome info change command to add personal information.
    3. Issue the mmcallhome proxy command to include a proxy if needed.
    4. Issue the mmcallhome group add or mmcallhome group auto command to create call home groups .
  • To enable call home once the call home package is installed and the groups are configured, issue the mmcallhome capability enable command.
For information on tip events, see Event type and monitoring status for system health.
Note: Since the TIP state is only checked once every hour, it might take up to an hour for the change to reflect in the output of the mmhealth command.