Dennis, You seem to be running plock_rate_limit=100 that limits the number of plocks/sec to 100 to avoid network flooding due to plocks. Setting this as <gfs_controld plock_rate_limit="0"/> in cluster.conf should give you better plock performance. Hope this helps, Thanks! --Abhi Dennis B. Hopp wrote: > We have a three node nfs/samba cluster that we seem to be having very > poor performance on GFS2. We have a samba share that is acting as a > disk to disk backup share for Backup Exec and during the backup > process the load on the server will go through the roof until the > network requests timeout and the backup job fails. > > I downloaded the ping_pong utility and ran it and seem to be getting > terrible performance: > > [root@sc2 ~]# ./ping_ping /mnt/backup/test.dat 4 > 97 locks/sec > > The results are the same on all three nodes. > > I can't seem to figure out why this is so bad. Some additional information: > > [root@sc2 ~]# gfs2_tool gettune /mnt/backup > new_files_directio = 0 > new_files_jdata = 0 > quota_scale = 1.0000 (1, 1) > logd_secs = 1 > recoverd_secs = 60 > statfs_quantum = 30 > stall_secs = 600 > quota_cache_secs = 300 > quota_simul_sync = 64 > statfs_slow = 0 > complain_secs = 10 > max_readahead = 262144 > quota_quantum = 60 > quota_warn_period = 10 > jindex_refresh_secs = 60 > log_flush_secs = 60 > incore_log_blocks = 1024 > demote_secs = 600 > > [root@sc2 ~]# gfs2_tool getargs /mnt/backup > data 2 > suiddir 0 > quota 0 > posix_acl 1 > num_glockd 1 > upgrade 0 > debug 0 > localflocks 0 > localcaching 0 > ignore_local_fs 0 > spectator 0 > hostdata jid=0:id=262146:first=0 > locktable > lockproto lock_dlm > > 97 locks/sec > [root@sc2 ~]# rpm -qa | grep gfs > kmod-gfs-0.1.31-3.el5 > gfs-utils-0.1.18-1.el5 > gfs2-utils-0.1.53-1.el5_3.3 > > [root@sc2 ~]# uname -r > 2.6.18-128.1.10.el5 > > Thanks, > > --Dennis > > > -- > Linux-cluster mailing list > Linux-cluster@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/linux-cluster > > -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
