Attila Heidrich
2015-Nov-09 16:46 UTC
[Gluster-users] stat calls on gluster mounts lockups the system
Hi! I have been using glusterfs since years, mainly without severe problems. Now I have encountered a very strange behaviour, I cannot handle myself. Run Ubuntu 14.04s in AWS. Bricks are on EBS, replication peers are in different availability zones. In many cases this works like charm, no problam et all, but teher's one pair of inctances just cannot setup correctly. Peers are connected, everithing looks fine, I can reach the mounted resources, sync works two ways. The problem is, that applications like df, or even commandline substituion (pressingTAB) on the parent of the mount point locks up. Forever... I can reach my data,manipulate files, if I know the exact file names. This applies on dirs as well: root at control2:~# ls -ltr /mnt/glu-control-share/salt total 12 drwxrwxr-x 10 ubuntu ubuntu 4096 Jan 14 2015 xxxxxx-pillar lrwxrwxrwx 1 root root 29 Jan 14 2015 pillar -> xxxxxx-pillar/aws-pillar drwxrwxr-x 25 ubuntu ubuntu 4096 Jan 16 2015 xxxxxx-saltstack drwxrwxr-x 18 ubuntu ubuntu 4096 Jan 23 2015 states "ls /mnt" hangs forever mount is 127.0.0.1:/control-share on /mnt/glu-control-share type fuse.glusterfs (rw,default_permissions,allow_other,max_read=131072) fstab: 127.0.0.1:/control-share /mnt/glu-control-share glusterfs defaults,_netdev 0 0 heidricha at aldev-control1:~$ dpkg -l|grep gluster ii glusterfs-client 3.7.5-ubuntu1~trusty1 amd64 clustered file-system (client package) ii glusterfs-common 3.7.5-ubuntu1~trusty1 amd64 GlusterFS common libraries and translator modules ii glusterfs-server 3.7.5-ubuntu1~trusty1 amd64 clustered file-system (server package) heidricha at aldev-control1:~$ sudo gluster peer status [sudo] password for heidricha: Number of Peers: 1 Hostname: 172.16.2.10 Uuid: f6cc9444-6221-446d-be4e-4af9e93d15ba State: Peer in Cluster (Connected) heidricha at aldev-control1:~$ sudo gluster peer status Number of Peers: 1 Hostname: 172.16.2.10 Uuid: f6cc9444-6221-446d-be4e-4af9e93d15ba State: Peer in Cluster (Connected) heidricha at aldev-control1:~$ sudo gluster volume info Volume Name: control-share Type: Replicate Volume ID: ef80c184-4255-40a5-bfa3-faa0b69c334c Status: Started Number of Bricks: 1 x 2 = 2 Transport-type: tcp Bricks: Brick1: 172.16.1.10:/mnt/ebs-lv-data/shared Brick2: 172.16.2.10:/mnt/ebs-lv-data/shared Options Reconfigured: performance.readdir-ahead: on nfs.disable: on heidricha at aldev-control1:~$ sudo gluster volume status Status of volume: control-share Gluster process TCP Port RDMA Port Online Pid ------------------------------------------------------------------------------ Brick 172.16.1.10:/mnt/ebs-lv-data/shared 49152 0 Y 2771 Brick 172.16.2.10:/mnt/ebs-lv-data/shared 49152 0 Y 1690 Self-heal Daemon on localhost N/A N/A Y 2762 Self-heal Daemon on 172.16.2.10 N/A N/A Y 1682 Task Status of Volume control-share ------------------------------------------------------------------------------ There are no active volume tasks Where to look for the problem? Attila -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20151109/0f50c380/attachment.html>