Patrick Nixon
2019-Feb-12 02:58 UTC
[Gluster-users] Files on Brick not showing up in ls command
The files are being written to via the glusterfs mount (and read on the same client and a different client). I try not to do anything on the nodes directly because I understand that can cause weirdness. As far as I can tell, there haven't been any network disconnections, but I'll review the client log to see if there any indication. I don't recall any issues last time I was in there. Thanks for the response! On Mon, Feb 11, 2019 at 7:35 PM Vijay Bellur <vbellur at redhat.com> wrote:> > > On Sun, Feb 10, 2019 at 5:20 PM Patrick Nixon <pnixon at gmail.com> wrote: > >> Hello! >> >> I have an 8 node distribute volume setup. I have one node that accept >> files and stores them on disk, but when doing an ls, none of the files on >> that specific node are being returned. >> >> Can someone give some guidance on what should be the best place to start >> troubleshooting this? >> > > > Are the files being written from a glusterfs mount? If so, it might be > worth checking if the network connectivity is fine between the client (that > does ls) and the server/brick that contains these files. You could look up > the client log file to check if there are any messages related to > rpc disconnections. > > Regards, > Vijay > > >> # gluster volume info >> >> Volume Name: gfs >> Type: Distribute >> Volume ID: 44c8c4f1-2dfb-4c03-9bca-d1ae4f314a78 >> Status: Started >> Snapshot Count: 0 >> Number of Bricks: 8 >> Transport-type: tcp >> Bricks: >> Brick1: gfs01:/data/brick1/gv0 >> Brick2: gfs02:/data/brick1/gv0 >> Brick3: gfs03:/data/brick1/gv0 >> Brick4: gfs05:/data/brick1/gv0 >> Brick5: gfs06:/data/brick1/gv0 >> Brick6: gfs07:/data/brick1/gv0 >> Brick7: gfs08:/data/brick1/gv0 >> Brick8: gfs04:/data/brick1/gv0 >> Options Reconfigured: >> cluster.min-free-disk: 10% >> nfs.disable: on >> performance.readdir-ahead: on >> >> # gluster peer status >> Number of Peers: 7 >> Hostname: gfs03 >> Uuid: 4a2d4deb-f8dd-49fc-a2ab-74e39dc25e20 >> State: Peer in Cluster (Connected) >> Hostname: gfs08 >> Uuid: 17705b3a-ed6f-4123-8e2e-4dc5ab6d807d >> State: Peer in Cluster (Connected) >> Hostname: gfs07 >> Uuid: dd699f55-1a27-4e51-b864-b4600d630732 >> State: Peer in Cluster (Connected) >> Hostname: gfs06 >> Uuid: 8eb2a965-2c1e-4a64-b5b5-b7b7136ddede >> State: Peer in Cluster (Connected) >> Hostname: gfs04 >> Uuid: cd866191-f767-40d0-bf7b-81ca0bc032b7 >> State: Peer in Cluster (Connected) >> Hostname: gfs02 >> Uuid: 6864c6ac-6ff4-423a-ae3c-f5fd25621851 >> State: Peer in Cluster (Connected) >> Hostname: gfs05 >> Uuid: dcecb55a-87b8-4441-ab09-b52e485e5f62 >> State: Peer in Cluster (Connected) >> >> All gluster nodes are running glusterfs 4.0.2 >> The clients accessing the files are also running glusterfs 4.0.2 >> Both are Ubuntu >> >> Thanks! >> _______________________________________________ >> Gluster-users mailing list >> Gluster-users at gluster.org >> https://lists.gluster.org/mailman/listinfo/gluster-users > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20190211/66a7b443/attachment.html>
Nithya Balachandran
2019-Feb-14 03:28 UTC
[Gluster-users] Files on Brick not showing up in ls command
On Tue, 12 Feb 2019 at 08:30, Patrick Nixon <pnixon at gmail.com> wrote:> The files are being written to via the glusterfs mount (and read on the > same client and a different client). I try not to do anything on the nodes > directly because I understand that can cause weirdness. As far as I can > tell, there haven't been any network disconnections, but I'll review the > client log to see if there any indication. I don't recall any issues last > time I was in there. > >If I understand correctly, the files are written to the volume from the client , but when the same client tries to list them again, those entries are not listed. Is that right? Do the files exist on the bricks? Would you be willing to provide a tcpdump of the client when doing this? If yes, please do the following: On the client system: - tcpdump -i any -s 0 -w /var/tmp/dirls.pcap tcp and not port 22 - Copy the files to the volume using the client - List the contents of the directory in which the files should exist - Stop the tcpdump capture and send it to us. Also provide the name of the directory and the missing files. Regards, NIthya> Thanks for the response! > > On Mon, Feb 11, 2019 at 7:35 PM Vijay Bellur <vbellur at redhat.com> wrote: > >> >> >> On Sun, Feb 10, 2019 at 5:20 PM Patrick Nixon <pnixon at gmail.com> wrote: >> >>> Hello! >>> >>> I have an 8 node distribute volume setup. I have one node that accept >>> files and stores them on disk, but when doing an ls, none of the files on >>> that specific node are being returned. >>> >>> Can someone give some guidance on what should be the best place to >>> start troubleshooting this? >>> >> >> >> Are the files being written from a glusterfs mount? If so, it might be >> worth checking if the network connectivity is fine between the client (that >> does ls) and the server/brick that contains these files. You could look up >> the client log file to check if there are any messages related to >> rpc disconnections. >> >> Regards, >> Vijay >> >> >>> # gluster volume info >>> >>> Volume Name: gfs >>> Type: Distribute >>> Volume ID: 44c8c4f1-2dfb-4c03-9bca-d1ae4f314a78 >>> Status: Started >>> Snapshot Count: 0 >>> Number of Bricks: 8 >>> Transport-type: tcp >>> Bricks: >>> Brick1: gfs01:/data/brick1/gv0 >>> Brick2: gfs02:/data/brick1/gv0 >>> Brick3: gfs03:/data/brick1/gv0 >>> Brick4: gfs05:/data/brick1/gv0 >>> Brick5: gfs06:/data/brick1/gv0 >>> Brick6: gfs07:/data/brick1/gv0 >>> Brick7: gfs08:/data/brick1/gv0 >>> Brick8: gfs04:/data/brick1/gv0 >>> Options Reconfigured: >>> cluster.min-free-disk: 10% >>> nfs.disable: on >>> performance.readdir-ahead: on >>> >>> # gluster peer status >>> Number of Peers: 7 >>> Hostname: gfs03 >>> Uuid: 4a2d4deb-f8dd-49fc-a2ab-74e39dc25e20 >>> State: Peer in Cluster (Connected) >>> Hostname: gfs08 >>> Uuid: 17705b3a-ed6f-4123-8e2e-4dc5ab6d807d >>> State: Peer in Cluster (Connected) >>> Hostname: gfs07 >>> Uuid: dd699f55-1a27-4e51-b864-b4600d630732 >>> State: Peer in Cluster (Connected) >>> Hostname: gfs06 >>> Uuid: 8eb2a965-2c1e-4a64-b5b5-b7b7136ddede >>> State: Peer in Cluster (Connected) >>> Hostname: gfs04 >>> Uuid: cd866191-f767-40d0-bf7b-81ca0bc032b7 >>> State: Peer in Cluster (Connected) >>> Hostname: gfs02 >>> Uuid: 6864c6ac-6ff4-423a-ae3c-f5fd25621851 >>> State: Peer in Cluster (Connected) >>> Hostname: gfs05 >>> Uuid: dcecb55a-87b8-4441-ab09-b52e485e5f62 >>> State: Peer in Cluster (Connected) >>> >>> All gluster nodes are running glusterfs 4.0.2 >>> The clients accessing the files are also running glusterfs 4.0.2 >>> Both are Ubuntu >>> >>> Thanks! >>> _______________________________________________ >>> Gluster-users mailing list >>> Gluster-users at gluster.org >>> https://lists.gluster.org/mailman/listinfo/gluster-users >> >> _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > https://lists.gluster.org/mailman/listinfo/gluster-users-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20190214/4502d258/attachment-0001.html>