I have run into this exact issue before and resolved it by simply syncing over the missing files and running a heal on the volume (can take a little time to correct) On Fri, Jan 31, 2020 at 7:05 PM Christian Reiss <email(a)christian-reiss.de&gt; wrote: > Hey folks, > > in our production setup with 3 nodes (HCI) we took one host down > (maintenance, stop gluster, poweroff via ssh/ovirt engine). Once it was > up the gluster hat 2k healing entries that went down in a matter on 10 > minutes to 2. > > Those two give me a headache: > > [root@node03:~] # gluster vol heal ssd_storage info > Brick node01:/gluster_bricks/ssd_storage/ssd_storage > <gfid:a121e4fb-0984-4e41-94d7-8f0c4f87f4b6> > <gfid:6f8817dc-3d92-46bf-aa65-a5d23f97490e> > Status: Connected > Number of entries: 2 > > Brick node02:/gluster_bricks/ssd_storage/ssd_storage > Status: Connected > Number of entries: 0 > > Brick node03:/gluster_bricks/ssd_storage/ssd_storage > <gfid:a121e4fb-0984-4e41-94d7-8f0c4f87f4b6> > <gfid:6f8817dc-3d92-46bf-aa65-a5d23f97490e> > Status: Connected > Number of entries: 2 > > No paths, only gfid. We took down node2, so it does not have the file: > > [root@node01:~] # md5sum > > /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > 75c4941683b7eabc223fc9d5f022a77c > > /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > > [root@node02:~] # md5sum > > /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > md5sum: > /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6: > > No such file or directory > > [root@node03:~] # md5sum > > /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > 75c4941683b7eabc223fc9d5f022a77c > > /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > > The other two files are md5-identical. > > These flags are identical, too: > > [root@node01:~] # getfattr -d -m . -e hex > > /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > getfattr: Removing leading '/' from absolute path names > # file: > > gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > > security.selinux=0x73797374656d5f753a6f626a6563745f723a676c7573746572645f627269636b5f743a733000 > trusted.afr.dirty=0x000000000000000000000000 > trusted.afr.ssd_storage-client-1=0x0000004f0000000100000000 > trusted.gfid=0xa121e4fb09844e4194d78f0c4f87f4b6 > > trusted.gfid2path.d4cf876a215b173f=0x62653331383633382d653861302d346336642d393737642d3761393337616138343830362f38366461303238392d663734662d343230302d393238342d3637386537626437363139352e31323030 > > trusted.glusterfs.mdata=0x010000000000000000000000005e349b1e000000001139aa2a000000005e349b1e000000001139aa2a000000005e34994900000000304a5eb2 > > getfattr: Removing leading '/' from absolute path names > # file: > > gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 > > security.selinux=0x73797374656d5f753a6f626a6563745f723a676c7573746572645f627269636b5f743a733000 > trusted.afr.dirty=0x000000000000000000000000 > trusted.afr.ssd_storage-client-1=0x0000004f0000000100000000 > trusted.gfid=0xa121e4fb09844e4194d78f0c4f87f4b6 > > trusted.gfid2path.d4cf876a215b173f=0x62653331383633382d653861302d346336642d393737642d3761393337616138343830362f38366461303238392d663734662d343230302d393238342d3637386537626437363139352e31323030 > > trusted.glusterfs.mdata=0x010000000000000000000000005e349b1e000000001139aa2a000000005e349b1e000000001139aa2a000000005e34994900000000304a5eb2 > > Now, I dont dare simply proceeding withouth some advice. > Anyone got a clue on who to resolve this issue? File #2 is identical to > this one, from a problem point of view. > > Have a great weekend! > -Chris. > > -- > with kind regards, > mit freundlichen Gruessen, > > Christian Reiss > _______________________________________________ > Users mailing list -- users(a)ovirt.org > To unsubscribe send an email to users-leave(a)ovirt.org > Privacy Statement: https://www.ovirt.org/site/privacy-policy/ > oVirt Code of Conduct: > https://www.ovirt.org/community/about/community-guidelines/ > List Archives: > https://lists.ovirt.org/archives/list/users@ovirt.org/message/FGIQFIRC6QY... >

Hey Strahil, thanks for your answer. On 01/02/2020 08:18, Strahil Nikolov wrote: > There is an active thread in gluster-users , so it will be nice to mention this there. > > About the sync, you can find the paths via: > 1. Mount > mount -t glusterfs -o aux-gfid-mount vm1:test /mnt/testvol > 2. Find the path of files: > getfattr -n trusted.glusterfs.pathinfo -e text/mnt/testvol/.gfid/<GFID> > > I bet it's the same file that causes me problems. Just verify the contents and you will see that one of them is newer -> just rsync it to the bricks node01 & node03 and run 'gluster volume heal <volume> full.' I did a cross-post to gluster-users just now. You are right, the brick-files have a slightly different timestamp: [root@node01:~] # stat /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 File: ‘/gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6’ Size: 67108864 Blocks: 54576 IO Block: 4096 regular file Device: fd09h/64777d Inode: 16152829909 Links: 2 Access: (0660/-rw-rw----) Uid: ( 0/ root) Gid: ( 0/ root) Context: system_u:object_r:glusterd_brick_t:s0 Access: 2020-01-31 22:16:57.812620635 +0100 Modify: 2020-02-01 07:19:24.183045141 +0100 Change: 2020-02-01 07:19:24.186045203 +0100 Birth: - [root@node03:~] # stat /gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6 File: ‘/gluster_bricks/ssd_storage/ssd_storage/.glusterfs/a1/21/a121e4fb-0984-4e41-94d7-8f0c4f87f4b6’ Size: 67108864 Blocks: 54576 IO Block: 4096 regular file Device: fd09h/64777d Inode: 16154259424 Links: 2 Access: (0660/-rw-rw----) Uid: ( 0/ root) Gid: ( 0/ root) Context: system_u:object_r:glusterd_brick_t:s0 Access: 2020-01-31 22:16:57.811800217 +0100 Modify: 2020-02-01 07:19:24.180939487 +0100 Change: 2020-02-01 07:19:24.184939586 +0100 Birth: - Contents (getfattr, md5) are still identical. I am unsure about your suggested rsync, tho: - node1 has file, - node2 does not, - node3 has file. So I can rsync node1 to node2 and node3 or node3 to node1 and node2. Sync to node1 and node3 cant be done as node2 does not have the file. Can I do the rsync on a live, running Gluster?