Description of problem: Linux untar followed by rm -rf * on Fuse sub-dir mount results in "rm: fts_read failed" Version-Release number of selected component (if applicable): glusterfs-fuse-3.8.4-49.el7rhgs.x86_64 How reproducible: 2/2 Steps to Reproduce: 1.Create 4*3 Distributed-Replicate Volume.Enable quota on volume 2.Mount the volume on client via fuse 3.Create 10 directories inside the mount point 4.Mount the sub-dir on same client on different mount points 5.Copy the linux tar file on all 10 sub-dir mount points 6.Perform linux untar on all the 10 sub-dir mount point simultaneously. 7.After linux untar is completed,perform rm -rf * on any 1 sub-dir mount point # cat sub-dirs-create.sh mkdir /mnt/mani-subdir-test mount -t glusterfs dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1 /mnt/mani-subdir-test for i in $(seq 1 10) do mkdir /mnt/mani-subdir-test/dir$i mkdir /mnt/Pdir$i mount -t glusterfs dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir$i /mnt/Pdir$i sleep 2 done for i in $(seq 1 10) do cp /home/linux-4.9.5.tar.xz /mnt/Pdir$i cd /mnt/Pdir$i echo "tar xvf /mnt/Pdir$i/linux-4.9.5.tar.xz" > file.sh nohup sh file.sh & done Actual results: rm -rf * fails on Fuse sub-sir mount.Tried the same on other sub-dirs mounted as well,none was successfull # rm -rf * rm: fts_read failed: No such file or directory Expected results: rm -rf * should not fail Additional info: Tried copying the same linux tar file on volume mount point(fuse).Perform rm -rf *,files got removed successfully # df -hT Filesystem Type Size Used Avail Use% Mounted on /dev/mapper/rhel_dhcp37--142-root xfs 17G 2.2G 15G 13% / devtmpfs devtmpfs 1.9G 0 1.9G 0% /dev tmpfs tmpfs 1.9G 0 1.9G 0% /dev/shm tmpfs tmpfs 1.9G 57M 1.8G 3% /run tmpfs tmpfs 1.9G 0 1.9G 0% /sys/fs/cgroup /dev/sda1 xfs 1014M 214M 801M 22% /boot tmpfs tmpfs 380M 0 380M 0% /run/user/0 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1 fuse.glusterfs 80G 8.9G 71G 12% /mnt/mani-subdir-test dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir2 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir2 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir1 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir1 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir3 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir3 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir4 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir4 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir5 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir5 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir6 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir6 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir7 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir7 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir8 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir8 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir9 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir9 dhcp42-125.lab.eng.blr.redhat.com:ganeshavol1/dir10 fuse.glusterfs 80G 8.9G 71G 12% /mnt/Pdir10 Attaching sosreports shortly
Can you point me to gluster logs?
I don't think this has got anything to do with subdir mount. I hit this some time back as well : https://bugzilla.redhat.com/show_bug.cgi?id=1439599
I am in the process of reproducing the bug, does the bug happen if we don't have quota? > 1.Create 4*3 Distributed-Replicate Volume.Enable quota on volume Instead just have a 4*3 volume, but not enabling quota?
(In reply to Amar Tumballi from comment #5) > I am in the process of reproducing the bug, does the bug happen if we don't > have quota? > > > 1.Create 4*3 Distributed-Replicate Volume.Enable quota on volume > > Instead just have a 4*3 volume, but not enabling quota? Hey Amar, The issue is reproducible with quota disable as well.
Raising the blocker flag for 3.3.1 as the issue is consistent (with same use case) on Fuse sub-dir Mount.
Noticed that with fix posted for bz1505473, even this bz will be resolved. I could reproduce the issue always before https://review.gluster.org/#/c/18558, with the fix, not able to reproduce this. Will move this to ON_QA once the patch gets merged in master, and then gets backported.
Upstream patch : https://review.gluster.org/#/c/18558
https://code.engineering.redhat.com/gerrit/121726
Verified this with glusterfs build glusterfs-3.8.4-51.el7rhgs.x86_64 rm -rf * after linux untar is performed on subdir mount, remove all files from mount point.Hence moving this bug to verified state.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2017:3276