[LU-5785] recovery-mds-scale test_failover_ost: test_failover_ost returned 1

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.7.0
• Type: Bug
• Priority: Minor
• Reporter: Maloo
• Assignee: Mikhail Pershin
• Resolution: Duplicate

Environment

client and server: lustre-master build # 2690
client is SLES11 SP3

• Severity: 3

Description

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

This issue relates to the following test suite run: https://testing.hpdd.intel.com/test_sets/548b249c-565a-11e4-97f1-5254006e85c2.

The sub-test test_failover_ost failed with the following error:

test_failover_ost returned 1

Lustre: DEBUG MARKER: ==== Checking the clients loads AFTER failover -- failure NOT OK
Lustre: DEBUG MARKER: /usr/sbin/lctl mark ost5 has failed over 1 times, and counting...
Lustre: DEBUG MARKER: ost5 has failed over 1 times, and counting...
Lustre: lustre-OST0006: Will be in recovery for at least 1:00, or until 3 clients reconnect
LustreError: 360:0:(ldlm_lib.c:1730:check_for_next_transno()) lustre-OST0006: waking for gap in transno, VBR is OFF (skip: 4295345303, ql: 2, comp: 1, conn: 3, next: 4295345312, last_committed: 4295345240)
Lustre: lustre-OST0006: Recovery over after 0:22, of 3 clients 3 recovered and 0 were evicted.
LustreError: 30883:0:(ldlm_lockd.c:1288:ldlm_handle_enqueue0()) ### delayed lvb init failed (rc -2) ns: filter-lustre-OST0006_UUID lock: ffff88006744d080/0x8abfa2edda4a4ccc lrc: 2/0,0 mode: --/PR res: [0xde56:0x0:0x0].0 rrc: 1 type: EXT [0->0] (req 0->0) flags: 0x40000000000000 nid: local remote: 0x7a6f60280c307e9a expref: -99 pid: 30883 timeout: 0 lvb_type: 0
LustreError: 30888:0:(ofd_io.c:600:ofd_preprw_write()) lustre-OST0006: BRW to missing obj 0x0:56912
LustreError: 30888:0:(ofd_io.c:600:ofd_preprw_write()) Skipped 1 previous similar message
Lustre: DEBUG MARKER: /usr/sbin/lctl mark Duration:               86400

Issue Links

• duplicates: LU-4621 recovery-mds-scale: test_failover_ost
• is duplicated by: LU-5485 first mount always fail with avoid_as...
• is related to: LU-5782 recovery-mds-scale test_failover_ost:

Comments

[Jodi Levi (Inactive)]

Mike,
Could you please have a look at this one and comment?
Thank you!

[Andreas Dilger]

The console logs reports:

23:57:55:Lustre: lustre-OST0006: deleting orphan objects from 0x0:56878 to 0x0:56897
00:19:16:Lustre: lustre-OST0006: deleting orphan objects from 0x0:57116 to 0x0:57142

so it doesn't look like orphan cleanup caused the object to be removed. Maybe there is something in the debug logs?

I wonder if it makes sense to dump the debug log to shared storage just before the reboot to try and capture more useful information for debugging?

It looks like this test is run several times a day, but almost all of them are failing for one reason or another. It makes sense to investigate and fix some of these failures.

[Jian Yu]

On Client 3 (onyx-45vm6):

    tar: etc/sysconfig/network/ifroute-lo: Cannot write: No such file or directory
    tar: etc/sysconfig/network/routes: Cannot write: No such file or directory
    tar: etc/sysconfig/network/if-down.d/ndp-proxy: Cannot stat: No such file or directory
    tar: Exiting with failure status due to previous errors

It looks like this is a duplicate of LU-4621.

Hi Hongchao,
Could you please confirm this?