CONTAP-193448: Slow mroot's unmount and mgwd restarting occurs when reverting to 9.1x.1
Issue
- Along with system node revert-to command, cluster config rever occurs
- The mroot unmount and mgwd restarting also follow as a part of the revert process.
- The mroot unmount and mgwd restarting can take sufficiently long enough so that cluster shell fails to automatically re-connect and ONTAP drops to system shell.
Example:
----------------------------------------------------------------------------------------------------
::*> system node revert-to -version 9.11
The management application is restarting as part of the revert process. Log in again to complete the revert task.
The management gateway server restarted. Waiting to see if the connection can be reestablishedDec 12 12:33:43 [node1:spm.mgwd.process.exit:EMERGENCY]: Management Gateway (mgwd) subsystem with ID 2636 exited as a result of signal normal exit (0). The subsystem will attempt to restart.
The management gateway server restarted. Waiting to see if the connection can be reestablishedDec 12 12:38:00 [node1:monitor.globalStatus.critical:EMERGENCY]: Controller failover of node1 is not possible: downloading shelf firmware.
The management gateway server restarted. Waiting to see if the connection can be reestablished...
The connection with the management gateway server could not be reestablished.
Dec 12 12:39:00 [node1:monitor.globalStatus.critical:EMERGENCY]: Controller failover of node1 is not possible: Controller Failover takeover disabled.
umount: unmount of /mroot/etc/.usercores failed:
Device busy
The most recent coredump file is:
/mroot/etc/crash/
For further assistance, contact technical support.
Do you want to login to the systemshell as the 'diag' user? {yes, no}: umount: unmount of /mroot failed: Device busy
root 59059 - - mroot unmounted
Systemshell connected. Type in your diagnostic password to continue.
(diag@127.0.0.1) Password:
Warning: The system shell provides access to low-level
diagnostic tools that can cause irreparable damage to
the system if not used properly. Use this environment
only when directed to do so by support personnel.
node1%
----------------------------------------------------------------------------------------------------