Noel Dawe
2011-Mar-21 23:25 UTC
[Gluster-users] Gluster 3.1.3 rm: cannot remove `foo.bar' : Invalid argument
I cannot rm certain files and I see this in global.log: [2011-03-21 16:12:50.317110] W [fuse-bridge.c:905:fuse_unlink_cbk] 0-glusterfs-fuse: 686702: UNLINK() /endw/data/rel16/large/user.NoelDawe.00155569/optimized.group10.perf-tau.10404_003422.StreamD3PD_TauMEDIUM._00010.root => -1 (Invalid argument) Is there a known workaround other than manually deleting the files on each node? All the clocks are synced, if that eliminates anything that could cause this error. Here is the current config: # gluster volume info Volume Name: global Type: Distribute Status: Started Number of Bricks: 9 Transport-type: tcp Bricks: Brick1: lhc06:/scratch/export Brick2: lhc07:/scratch/export Brick3: lhc08:/scratch/export Brick4: lhc09:/scratch/export Brick5: lhc01:/scratch/export Brick6: lhc02:/scratch/export Brick7: lhc03:/scratch/export Brick8: lhc04:/scratch/export Brick9: lhc05:/scratch/export Options Reconfigured: performance.cache-size: 512MB performance.cache-refresh-timeout: 10 performance.write-behind-window-size: 64MB performance.read-ahead: on performance.write-behind: on performance.io-cache: on performance.stat-prefetch: on performance.quick-read: on performance.io-thread-count: 24 performance.flush-behind: on Thanks, Noel
Nic von Waltsleben
2011-Mar-23 13:13 UTC
[Gluster-users] Gluster 3.1.3 rm: cannot remove `foo.bar' : Invalid argument
Hi On 22 Mar 2011, at 1:25 AM, Noel Dawe wrote:> I cannot rm certain files and I see this in global.log: > > [2011-03-21 16:12:50.317110] W [fuse-bridge.c:905:fuse_unlink_cbk] > 0-glusterfs-fuse: 686702: UNLINK() > /endw/data/rel16/large/user.NoelDawe.00155569/optimized.group10.perf-tau.10404_003422.StreamD3PD_TauMEDIUM._00010.root > => -1 (Invalid argument) > > Is there a known workaround other than manually deleting the files on each > node? All the clocks are synced, if that eliminates anything that could > cause this error. >I believe we had the same problem a while back on 3.2.1, after poking around the internet with Google I ran the following commands on one of our Gluster hosts. It solved the problem for us but YMMV gluster volume set backup-volume performance.quick-read off gluster volume set backup-volume performance.stat-prefetch 0 You might also want to check the seemingly incomplete documentation to make sure this won't have some other detrimental impact to your GlusterFS. http://www.gluster.com/community/documentation/index.php/Gluster_3.1:_Setting_Volume_Options and the now outdated http://www.gluster.com/community/documentation/index.php/Translators Regards, Nic von Waltsleben nic at clue.co.za