Please don't reply to lustre-devel. Instead, comment in Bugzilla by using the
following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10960
What |Removed |Added
----------------------------------------------------------------------------
Attachment #9207 is|0 |1
obsolete| |
Created an attachment (id=9354)
Please don't reply to lustre-devel. Instead, comment in Bugzilla by using the
following link:
--> (https://bugzilla.lustre.org/attachment.cgi?id=9354&action=view)
ior-survey script
I got some errors while doing reread in the IOR test. On investigations, I
found
that the errors were due to bugs in ior_survey script, which surfaced after I
changed my cluster configuration. The synchronization between the two clients
went wrong due to difference between the I/O speeds of the two clients.
There is the speed difference between two nodes on which i am running
the ior_survey. The speed on rhea1 was 67.97 MB/sec whereas on rhea2 is
45.95 MB/sec. This was because when the last test which is reread is
completed the testfile gets deleted and in our case when one of the
nodes completes the reread the file gets deleted while the other tries
to read. So, the other node i.e. rhea2 cannot stat the testfile.
I have fixed the problem in the script by preventing the testfile from
deletion.
_______________________________________________
Lustre-devel mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-devel
