New Ibrix Clients Report "Segment.Rejected" Errors and Alerts

New Ibrix Clients Report "segment.rejected" Errors and
Alerts
Solution ID: 00000080
Created: 10/27/2008
DETAILS:
Symptoms:
Customer was receiving segment rejected alerts and seg.rejected errors in
/var/log/messages also from a handful of his clients.
Here is a sample error:
Segment.rejected:d9eeb202-36b0-4910-9889-beb67a210ec7, 1|
clientname, date-time
Cause:
<Cause>
Resolution:
When a segment is reported as rejected from a client[s] the Ibrix client is unable
to get to a particular portion/segment of the FS. There are a few checks involved
with troubleshooting this type of issue.
1. Validate from the segment server owner with lvmdiskscan, pvscan and
lvscan that the underlying device[s] are all active and superblk is readable
i.e. tuneibfs -l /dev/vg#/lv# ..
2. If the server does not have issue with mounting the segment move your
attention back to the clients.
3. On the clients experiencing the rejected.segment error; verify the
generation # of the FS matches that which is current on the servers and
FM.
i.e. ibrix_fs -i -f FSNAME
4. output from FM will show generation # also the compare
/usr/local/ibrix/bin/rtool enumfs outputs on the client and servers. Do they
match? In this case they did not and that meant the client[s] did not have
an updated map. Stale maps can cause segment access problems.

Summary of content (2 pages)