[LU-6248] sanity-scrub: Failed mount OST

• Type: Bug
• Resolution: Duplicate
• Priority: Minor
• Affects Version/s: Lustre 2.7.0, Lustre 2.8.0
• Labels: zfs
• Environment: client and server: lustre-master build # 2856; server is zfs
• Severity: 3
• Assignee: WC Triage
• Reporter: Maloo

Description

This issue was created by maloo for sarah <[e-mail]>

Please provide additional information about the failure here.

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/abe0dac2-b2d8-11e4-b42d-5254006e85c2.

Cannot find any log except this

CMD: onyx-45vm4 mkdir -p /mnt/ost1; mount -t lustre   		                   lustre-ost1/ost1 /mnt/ost1
onyx-45vm4: mount.lustre: mount lustre-ost1/ost1 at /mnt/ost1 failed: Cannot send after transport endpoint shutdown

Issue Links

• duplicates: LU-7193 sanity-scrub: No sub tests failed in this test set

Comments

[James Nunez (Inactive)]

Closing as a duplicate of LU-7193.

[James Nunez (Inactive)]

Note that the instances that Saurab listed above are all using ldiskfs, not zfs, and, as noted, they are all interop issues. Should we open a new ticket for the ldiskfs/interop failures?

[Saurabh Tandan (Inactive)]

Another instance found for interop : EL6.7 Server/2.5.5 Client
Server: master, build# 3303, RHEL 6.7
Client: 2.5.5, b2_5_fe/62
https://testing.hpdd.intel.com/test_sets/2af59388-bad6-11e5-9137-5254006e85c2

[Saurabh Tandan (Inactive)]

Another instance found for interop : EL6.7 Server/2.7.1 Client
Server: master, build# 3303, RHEL 6.7
Client: 2.7.1, b2_7_fe/34
https://testing.hpdd.intel.com/test_sets/3a41f816-bad8-11e5-87b4-5254006e85c2

[Saurabh Tandan (Inactive)]

Another instance for interop: EL7 Server/2.7.1 Client
Server: master, build# 3303, RHEL 7
Client: 2.7.1, b2_7_fe/34
https://testing.hpdd.intel.com/test_sets/606aa416-bac4-11e5-9137-5254006e85c2

[nasf (Inactive)]

Generally, we can check the fstype some early, but even if we can skip the sanity-scrub for zfs-based backend in time, we will hit the "Failed mount OST" in some subsequent test. Because the failure is not caused by OI scrub via same fail_loc injection (in sanity-scrub.sh) or OI scrub wrong repairing, instead, there should have been some trouble before the OI scrub.

[Isaac Huang (Inactive)]

Saw it once here:
https://testing.hpdd.intel.com/test_sets/784d3a32-b2a3-11e4-a8f5-5254006e85c2

Happened before the test was SKIPPED. Then I ran the test over 20 times but couldn't reproduce.

[Oleg Drokin]

Actually it appears we are testing this a bit late:

# build up a clean test environment.
formatall
setupall

[ $(facet_fstype $SINGLEMDS) != "ldiskfs" ] &&
        skip "test OI scrub only for ldiskfs" && check_and_cleanup_lustre &&
        exit 0

Should we move the tests before the reformat and remount since for ZFS apparently too small size might be unsuitable or something?

[nasf (Inactive)]

According to the test scripts, sanity-scrub will be skipped for zfs based backend.

[ $(facet_fstype $SINGLEMDS) != "ldiskfs" ] &&
        skip "test OI scrub only for ldiskfs" && check_and_cleanup_lustre &&
        exit 0
[ $(facet_fstype ost1) != "ldiskfs" ] &&
        skip "test OI scrub only for ldiskfs" && check_and_cleanup_lustre &&
        exit 0