[LU-8816] test-framework needs to reload SPL module after set hostid

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.9.0
• Severity: 3
• Assignee: Nathaniel Clark
• Reporter: Minh Diep

Description

recent failover testing shown failure

https://testing.hpdd.intel.com/test_sets/9560d0c2-a174-11e6-a031-5254006e85c2
https://testing.hpdd.intel.com/test_sets/f1680146-a58f-11e6-95cc-5254006e85c2

Issue Links

• is related to: LU-8694 ZFS format fails when /etc/hostid is set or unset.
• is related to: LU-7134 Ensure ZFS hostid protection if servicenode/failover options given to mkfs.lustre

Comments

[Minh Diep]

Peter,

Per Nate "There's an issue with ZFS code that doesn't call initial load of hostid (SPL::zone_gethostid()) until creation or importation of a pool."

So this is still an issue even after adding hostid in node provisioning

[Gerrit Updater]

Nathaniel Clark ([e-mail]) uploaded a new patch: http://review.whamcloud.com/23804
Subject: LU-8816 utils: Check /etc/hostid instead of failing for ZFS
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: a30b786793ff47cd65a07d6112a4e3cffbeb385a

[Peter Jones]

As I understand it, this issue is going to be handled in a DCO ticket

[Saurabh Tandan (Inactive)]

After Minh's comment above I retested and the testing could still not continue. Throwing the following error message:

11:45:05:onyx-37vm7: mkfs.lustre FATAL: spl_hostid not set. See mkfs.lustre(8)
11:45:05:onyx-37vm7: mkfs.lustre FATAL: mkfs failed 22
11:45:05:onyx-37vm7: mkfs.lustre: exiting with 22 (Invalid argument)
11:45:05:
11:45:05:   Permanent disk data:
11:45:05:Target:     lustre:MDT0000
11:45:05:Index:      0
11:45:05:Lustre FS:  lustre
11:45:05:Mount type: zfs
11:45:05:Flags:      0x65
11:45:05:              (MDT MGS first_time update )
11:45:05:Persistent mount opts: 
11:45:05:Parameters: failover.node=10.2.4.158@tcp sys.timeout=20 lov.stripesize=1048576 lov.stripecount=0 mdt.identity_upcall=/usr/sbin/l_getidentity
11:45:05:

Result - https://testing.hpdd.intel.com/test_sets/e141da18-ab74-11e6-986b-5254006e85c2

[Nathaniel Clark]

There's an issue with ZFS code that doesn't call initial load of hostid (SPL::zone_gethostid()) until creation or importation of a pool.

[Gerrit Updater]

Minh Diep ([e-mail]) uploaded a new patch: http://review.whamcloud.com/23684
Subject: LU-8816 test: reload SPL module after set_hostid
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 9e93a6a219ac22af49449ea147cb25d1af53c4c8