Red Hat Bugzilla – Bug 828701
rename failures with multiple clients
Last modified: 2015-09-01 19:05:55 EDT
Description of problem: With 2 fuse clients, trying to rename a directory from 2nd client and then renaming it back to the old name from the 1st client fails.
mkdir -p /mnt/gluster1/h1;
mv /mnt/gluster2/h1 /mnt/gluster2/h2;
mv /mnt/gluster1/h2 /mnt/gluster/h1;
Ran this script in a loop, and was able to reproduce it easily
Version-Release number of selected component (if applicable): 3.3.0qa45
How reproducible: Often
Rename fails with ENOENT
Shouldn't error out
By mounting the fuse client with --entry-timeout=0 and --attribute-timeout=0 and setting the md-cache timeout to 0, this issue wasn't seen.
Avati any update on this? I would say this is expected behavior in case where the parameters 'entry-timeout' 'attribute-timeout' and md-cache xlator are present... Let me know what you think about this?
Yes, this is expected behavior. Let's close this as not a bug.
*** Bug 852577 has been marked as a duplicate of this bug. ***
*** Bug 858486 has been marked as a duplicate of this bug. ***