Hello,
A few months ago, I enabled changelogs and registered a changelog reader on a
Lustre 2.5.3.90 file system. But I didn't regularly run a Robinhood changelog
reader for a long time - changelog entries accumulated, and nothing was
draining them until fairly recently. So a large backlog of changelog ids has
built up over time. I've since started running the Robinhood changelog reader
from Robinhood 3.0-rc1, which has started draining them.
I've recently did an "lfs changelog_clear ufrc-MDT0000 cl1 0" on our MDS and
followed it with a scan. Here's the value of mdd.*.changelog_users after I did
the changelog_clear:
[root@ufrcmds1 ~]# lctl get_param mdd.*.changelog_users
mdd.ufrc-MDT0000.changelog_users=
current index: 2434010839
ID index
cl1 2434001066
cl2 0
After the scan completed, I restarted the changelog reader. Robinhood resumed
processing changelog entries where it left off (still far behind index
2434001066). What I would like to do is throw away all the unprocessed
changelog entries from before the scan started, and resume processing with
changelog entry 2434001066. I'm not sure of the best way to do that.
In the VARS table, I see ChangelogLastCommit_MDT0000 - can I just stop the
changelog reader, change this value to 2434001065, and start up again?
Thanks,
Craig Prescott
UF Research Computing
------------------------------------------------------------------------------
What NetFlow Analyzer can do for you? Monitors network bandwidth and traffic
patterns at an interface-level. Reveals which users, apps, and protocols are
consuming the most bandwidth. Provides multi-vendor support for NetFlow,
J-Flow, sFlow and other flows. Make informed decisions using capacity planning
reports.http://sdm.link/zohodev2dev
_______________________________________________
robinhood-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/robinhood-support
