Kingsley
2015-Jan-30 11:32 UTC
[Gluster-users] gluster 3.6.2 "no data available" error - selinux?
Hi, I've got a cluster running on CentOS 7 and Gluster 3.6.2 and using the fuse client on the clients, which are running the same OS and Gluster version as the servers. There are 4 servers with a brick each, running 4 way replication. On a client, each time I create a file or chmod it (it may happen with other operations but I haven't tested those), I get an error in the logs on the servers, although on the client the file operation appears to work ok. I get two lines like this: [2015-01-30 10:46:06.889661] E [marker.c:2562:marker_removexattr_cbk] 0-gv0-marker: No data available occurred while creating symlinks [2015-01-30 10:46:06.889685] I [server-rpc-fops.c:693:server_removexattr_cbk] 0-gv0-server: 29800868: REMOVEXATTR /testing/thingy (0460a967-d1de-462b-be97-d6e5736c89c4) of key security.ima ==> (No data available) The bricks are within /data on the servers; "mount" says this about /data on the servers: /dev/md123 on /data type xfs (rw,relatime,seclabel,attr2,inode64,sunit=1024,swidth=3072,noquota) This post says something about selinux being required in some circumstances: http://www.gluster.org/pipermail/gluster-users/2014-August/018387.html Do I need to set up selinux on the fuse clients? I've not really played with selinux before so would prefer to avoid that learning curve at this moment unless it's necessary. -- Cheers, Kingsley.
Kingsley
2015-Jan-30 15:59 UTC
[Gluster-users] gluster 3.6.2 "no data available" error - selinux?
Following on from this, I've put selinux into permissive mode and rebooted. The errors are still flooding the gluster logs but nothing is writing to /var/log/audit/audit.log so I can't see this being an selinux issue. Any clues? Any googling only leads me to suggestions about selinux. Cheres, Kingsley. On Fri, 2015-01-30 at 11:32 +0000, Kingsley wrote:> Hi, > > I've got a cluster running on CentOS 7 and Gluster 3.6.2 and using the > fuse client on the clients, which are running the same OS and Gluster > version as the servers. There are 4 servers with a brick each, running 4 > way replication. > > On a client, each time I create a file or chmod it (it may happen with > other operations but I haven't tested those), I get an error in the logs > on the servers, although on the client the file operation appears to > work ok. I get two lines like this: > > [2015-01-30 10:46:06.889661] E [marker.c:2562:marker_removexattr_cbk] 0-gv0-marker: No data available occurred while creating symlinks > [2015-01-30 10:46:06.889685] I [server-rpc-fops.c:693:server_removexattr_cbk] 0-gv0-server: 29800868: REMOVEXATTR /testing/thingy (0460a967-d1de-462b-be97-d6e5736c89c4) of key security.ima ==> (No data available) > > The bricks are within /data on the servers; "mount" says this > about /data on the servers: > > /dev/md123 on /data type xfs (rw,relatime,seclabel,attr2,inode64,sunit=1024,swidth=3072,noquota) > > This post says something about selinux being required in some > circumstances: > http://www.gluster.org/pipermail/gluster-users/2014-August/018387.html > > Do I need to set up selinux on the fuse clients? I've not really played > with selinux before so would prefer to avoid that learning curve at this > moment unless it's necessary. >
Vijay Bellur
2015-Feb-01 23:08 UTC
[Gluster-users] gluster 3.6.2 "no data available" error - selinux?
On 01/30/2015 04:59 PM, Kingsley wrote:> Following on from this, I've put selinux into permissive mode and > rebooted. The errors are still flooding the gluster logs but nothing is > writing to /var/log/audit/audit.log so I can't see this being an selinux > issue. > > Any clues? Any googling only leads me to suggestions about selinux. >These logs are benign, though excessive and hence annoying. The log level for "no data available" errors in the removexattr path have been lowered to DEBUG or TRACE in mainline so that they do not appear repetitively by default in the log files. I have backported those patches ([1] and [2]) to release-3.6 now. We should have this fixed in 3.6.3. Cheers, Vijay [1] http://review.gluster.org/9524 [2] http://review.gluster.org/9525> Cheres, > Kingsley. > > On Fri, 2015-01-30 at 11:32 +0000, Kingsley wrote: >> Hi, >> >> I've got a cluster running on CentOS 7 and Gluster 3.6.2 and using the >> fuse client on the clients, which are running the same OS and Gluster >> version as the servers. There are 4 servers with a brick each, running 4 >> way replication. >> >> On a client, each time I create a file or chmod it (it may happen with >> other operations but I haven't tested those), I get an error in the logs >> on the servers, although on the client the file operation appears to >> work ok. I get two lines like this: >> >> [2015-01-30 10:46:06.889661] E [marker.c:2562:marker_removexattr_cbk] 0-gv0-marker: No data available occurred while creating symlinks >> [2015-01-30 10:46:06.889685] I [server-rpc-fops.c:693:server_removexattr_cbk] 0-gv0-server: 29800868: REMOVEXATTR /testing/thingy (0460a967-d1de-462b-be97-d6e5736c89c4) of key security.ima ==> (No data available) >> >> The bricks are within /data on the servers; "mount" says this >> about /data on the servers: >> >> /dev/md123 on /data type xfs (rw,relatime,seclabel,attr2,inode64,sunit=1024,swidth=3072,noquota) >> >> This post says something about selinux being required in some >> circumstances: >> http://www.gluster.org/pipermail/gluster-users/2014-August/018387.html >> >> Do I need to set up selinux on the fuse clients? I've not really played >> with selinux before so would prefer to avoid that learning curve at this >> moment unless it's necessary. >> > > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://www.gluster.org/mailman/listinfo/gluster-users >