[Linux-cluster] GFS 6.0.24 hangs machine....
Corey Kovacs
cjkovacs at verizon.net
Sun Apr 17 11:01:35 UTC 2005
Hello,
I've got a 5 node GFS cluster (RHEL3u4, GFS 6.0.2-24, kernel 2.4.21-24.0.1)
with 3 volumes, one of which is approx 500GB and contains several thousand
small files. When I do a find on that volume, or slocate is run via it's cron
job, or I rsync that volume, the node used to access the volume get's into a
state where it cannot fork anymore and nothing can be done with the machine
until it is restarted (usually requiring a "fence_node" from another
machine). The cluster is configured with 3 of the nodes acting as lock
managers, using DL360's with 2GB ram each and qlogic 2342 dual port cards
connected to an msa1000. The journals are not on there own volumes and the
defaults are used for mounting.
Is this a known problem? I've searched for other posts with this problem but
have not had any luck with it. Any ideas as to what might be causing this?
Thanks
Corey
More information about the Linux-cluster
mailing list