Pranith Kumar Karampuri

Without stack trace or core file, it won't be possible to debug this issue. Please attach the relevant info.

@ikbalcharen .glusterfs directory stores hardlinks to the actual files on the bricks. So it will have the size closer to the brick usage. So please don't delete anything there.

@sbk173 This is the issue

@sanjurakonde To run the snapshot tests on LVM based setup and post the results again @pranithk to run DHT ones and report @rafikc30 to report the status on SLES15

distribute xlator results: ``` tests/basic/distribute/file-create.t (Wstat: 0 Tests: 32 Failed: 5) Failed tests: 6, 8, 10-11, 26 tests/basic/distribute/force-migration.t (Wstat: 0 Tests: 21 Failed: 1) Failed test: 16 Files=16, Tests=661, 187...

ta-read.t will take some time I think tests/basic/afr: - arbiter-mount.t - This is failing because nfs3 mount based testing - When I compile with --enable-gnfs, and starting rpc.statd before running...

@rafikc30 ``` root@glusterfs-fedora-41:/home/pk/glusterfs# prove -rfv ./tests/bugs/replicate/bug-1340623-mkdir-fails-remove-brick-started.t ./tests/bugs/replicate/bug-1340623-mkdir-fails-remove-brick-started.t .. egrep: warning: egrep is obsolescent; using grep -E egrep: warning: egrep is obsolescent; using grep -E 1..16 ok 1 [ 1800/ 1034]...

@rafikc30 Only the following test cases were failing on fedora 41 ``` Test Summary Report ------------------- ./tests/bugs/shard/unlinks-and-renames.t (Wstat: 0 Tests: 203 Failed: 12) Failed tests: 111-112, 143-144, 165-170, 174-175 ./tests/features/mandatory-lock-forced.t...

These were mostly fixed in 11.2 itself. Will create another issue when 11.3 release is almost there.

The issue seems to be that there is a lag between setting ctx->cleanup_started and setting reg->fin in timer, so by the time the event callback is triggered conn was already...