LLNL / lmt

Lustre Monitoring Tools
GNU General Public License v2.0
67 stars 21 forks source link

Report recovery status for MDT #27

Closed ofaaland closed 8 years ago

ofaaland commented 8 years ago

Update the MDT cerebro metric to report the recovery status, from the procfile recovery_status, as is already done for OSTs.

Move code to parse the procfile out of ost.c to enable it to be used by mdt.c as well.

Define a version 2 of the lmt_mdt message, which includes the recovery status field.

Define a parser for the lmt_mdt version 2 message, and modify ltop to use the appropriate version.

When an MDT reports it is in RECOVERY, display a message in ltop at the top of the window to the right of the filesystem name. If more than one MDT is in recovery, the full status of the first one encountered will be displayed. This allows the user to determine whether the count of re-connected clients is increasing or not.

morrone commented 8 years ago

Landed!