Description of problem:
While trying to reproduce https://bugzilla.redhat.com/show_bug.cgi?id=1529072 on a build older than the fix, it is found that if parallel-readdir is enabled, the readdir-optimize option even when it is set to on it is behaving as off.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
There are no specific steps to reproduce this issue. It can be reproduced by attaching gdb or by looking at tcpdump wireshark.
When parallel-readdir is enabled, the readdir-optimize option when set to on it is behaving as off.
When parallel-readdir is enabled, the readdir-optimize option when set to on should continue as on.
Patch posted Upstream: https://review.gluster.org/20026, and i think its hdoul be taken for 3.4 as its a regression that can disable readdir-optimize option.
Verified this BZ on glusterfs version 3.12.2-11.el7rhgs.x86_64 using wireshark.
Followed below steps,
1) Created a distributed volume and started it.
2) Enabled parallel-readdir and readdir-optimize.
3) FUSE mounted the volume on a client and created directories (lets say terminal-1).
4) From the same client another terminal (terminal-2), started collecting statedump using below command
"tcpdump -n -tttt -w glusterfs.pcap -i any -s 0 tcp"
5) from client terminal-1, run 'ls' command and once the ls command succeeds stop collecting statedump.
6) Analyse the statedumps using wireshark.
After analyzing the statedump using wireshark, it was confirmed that the requests were sent between first brick of the volume and the client which should be the expected behavior.
Hence, moving this BZ 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.