Triage and Data Collection

Triage and Data Collection

When you run into an issue with the NetApp OpenStack integrations, you may

OpenStack log and configuration files provide information that aid in triaging bugs. As a rule of thumb log files need to be collected for the processes that were involved. You may enable logging at the debug level and attempt to reproduce the issue to trace through the logs. The option to toggle debug logging is called debug in the respective configuration file. This option defaults to False and can be set to True when troubleshooting.

Be aware that debug logging will result in bloated log files and is typically turned off except when identifying root cause for failures or unexpected behavior. Also be aware of log rotation settings and attempt to collate archived log files as necessary.

If using Cinder, the following processes log into individual files:

  • cinder-api
  • cinder-backup
  • cinder-scheduler
  • cinder-volume

Note

When using ONTAP as your block storage back end, you can add the following lines to your NetApp backend stanza(s) in cinder.conf to capture more debug logging around the driver’s interaction with ONTAP in the cinder-volume log:

trace_flags = method,api

Note

To trace specific ONTAP APIs, add the following line to your NetApp backend stanza(s) in cinder.conf. The following line lists all APIs beginning with volume:

netapp_api_trace_pattern = ^volume-.*$

Note

To trace all ONTAP APIs with exception of those beginning with perf add the following line to your NetApp backend stanza(s) in cinder.conf:

netapp_api_trace_pattern = ^(?!(perf)).*$

If using Manila, the following processes log into individual files:

  • manila-api
  • manila-scheduler
  • manila-share
  • manila-data

Note

When using ONTAP as your shared filesystems storage back end, you can add the following lines to your NetApp backend stanza(s) in manila.conf to capture more debug logging around the driver’s interaction with ONTAP in the manila-share log:

netapp_trace_flags = method,api

Note

To trace specific ONTAP APIs, add the following line to your NetApp backend stanza(s) in manila.conf:

netapp_api_trace_pattern  =  ^volume-.*$

Note

To trace all ONTAP APIs with exception of those that start with perf, add the following line to your NetApp backend stanza(s) in manila.conf:

netapp_api_trace_pattern  =  ^(?!(perf)).*$

If using Nova, the following processes log into individual files:

  • nova-api
  • nova-scheduler
  • nova-cpu

If using Glance, the following processes log into individual files:

  • glance-api
  • glance-registry

If using Swift, the following processes log into individual files:

  • swift-object-server
  • swift-object-replicator
  • swift-object-updator
  • swift-object-auditor
  • swift-container-server
  • swift-container-replicator
  • swift-container-updator
  • swift-container-auditor
  • swift-account-server
  • swift-account-replicator
  • swift-account-auditor

Besides log files, a support engineer would ask you to provide your Configuration files, with any sensitive information removed as necessary. The default location of the configuration files are in /etc directory on the controller nodes running those processes. For example, the default configuration files for Cinder are in /etc/cinder/ directory.