[Gluster-users] Weird problem with Gluster 3.0.8

Jeremy Mann jeremy at biochem.uthscsa.edu
Mon Jan 24 18:11:38 UTC 2011


This weekend we discovered a weird problem with our glusterfs filesystem
(at least we think it is). If we run a hmmr search on a database contained
within glusterfs, that node in question crashes for no reason. This is not
a normal crash, but a hard crash and the node completely turns off. At
first I thought it was Infiniband so I disabled Infiniband and switched
the storage bricks to use the internal gigE network and reran the hmmr
search. Same thing, the node instantly turns off and I have to go across
the room to turn it back on. Not to kill anybody else's jobs, but I have
tried this on 3 separate nodes and they all turn off on this same search.

Interestingly, if we run the same search against the database contained
within his home NFS shared directory, it runs perfectly.

I am at a loss here, so does anybody have any insight or something I can
do to debug this weird problem?


-- 
Jeremy Mann
jeremy at biochem.uthscsa.edu

University of Texas Health Science Center
Bioinformatics Core Facility
http://www.bioinformatics.uthscsa.edu
Phone: (210) 767-3419
                        ^^^^^^^^
** NEW OFFICE NUMBER **




More information about the Gluster-users mailing list