Configuring debug data collection for NFS hung scenarios

You can configure Ganesha core dump and GPFS trace collection when NFS-Ganesha is detected as hung or unresponsive.

Before you begin

Make sure that the following requirements are met:

You have administrative access to the IBM Storage Scale environment.
You have sufficient space in /var/adm/ras/ and /tmp/ before you enable these options, and monitor disk usage during active troubleshooting sessions.

About this task

Additional diagnostic data can be collected when NFS-Ganesha becomes unresponsive while the service process remains active. The configuration file location is mmsysmonitor.conf.

These options control the collection of diagnostic information when the mmsysmon service detects a hung condition.

The collectnfscoreonhang and collectgpfstraceonhang options control what additional data is collected during this event, and you can enable or disable these options independently based on troubleshooting needs. By default the collectnfscoreonhang option is disabled and the collectgpfstraceonhang option is enabled.

Attention: Use these configuration options only under the guidance of IBM Storage Scale Support. Changes to these settings can affect performance, resource usage, diagnostic data collection, and normal system operation. Retain the default settings unless IBM Support requests a change for troubleshooting.

Procedure

Check whether the configuration options are enabled or disabled.
```
mmlsconfig |grep mmhealth
```
Note: Use the mmchconfig command to update configuration options without restarting mmsysmon service.
Enable core dump collection on hang when you want to collect a Ganesha core dump when NFS is detected as hung or unresponsive.
```
mmchconfig mmhealth-nfs-collectnfscoreonhang=yes --force
```
When the mmhealth-nfs-collectnfscoreonhang option enabled, the hang event collection can include network statistics, Ganesha statistics, cluster-wide waiter information, GPFS statistics, process limits, stack traces, and the core dump through gcore.
Disable core dump collection on hang when you do not need the core dump.
```
mmchconfig mmhealth-nfs-collectnfscoreonhang=no --force
```
Enable GPFS trace collection on hang when you want to collect GPFS trace data for deadlock and internal state analysis.
```
mmchconfig mmhealth-nfs-collectgpfstraceonhang=yes --force
```
Disable GPFS trace collection on hang when you no longer need it.
```
mmchconfig mmhealth-nfs-collectgpfstraceonhang=no --force
```

Results

Your CES NFS environment collects additional diagnostic data when the system detects that NFS-Ganesha is unresponsive or hung.