[LU-8694] ZFS format fails when /etc/hostid is set or unset.

• Type: Bug
• Resolution: Fixed
• Priority: Critical
• Fix Version/s: Lustre 2.9.0
• Affects Version/s: Lustre 2.9.0
• Severity: 3
• Assignee: Nathaniel Clark
• Reporter: Cliff White (Inactive)

Description

Attempting to format ZFS OSTs we see the following error:
mkfs.lustre FATAL: spl_hostid not set. See mkfs.lustre(8)

We attempt to set hostid by using /sbin/genhostid, we confirm the ID exists, same failure. We attempt to set hostid by /usr/bin/hostid, confirm the ID exists, same failure
At this time we cannot format OSTS, system is dead.

Issue Links

• is related to: LU-8816 test-framework needs to reload SPL module after set hostid
• is related to: LU-7314 In kiblnd_rejected(), NULL pointer 'cp' may be passed to function and can be dereferenced there

Comments

[Peter Jones]

Landed for 2.9

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch http://review.whamcloud.com/23118/
Subject: LU-8694 docs: ZFS hostid mkfs.lustre(8) man page update
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 5de18d44e1074d18f72946b2a19cd1b3a0b749c8

[Gerrit Updater]

Nathaniel Clark ([e-mail]) uploaded a new patch: http://review.whamcloud.com/23118
Subject: LU-8694 docs: ZFS hostid mkfs.lustre(8) man page update
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: b3e16b15a405d3993bdd8d0938977f294a670bc6

[Andreas Dilger]

This mostly means that we should change the mkfs.lustre message to indicate the spl.ko module needs to be reloaded (or better yet, try to unload it and reload it internally once), and improve the mkfs.lustre.8 man page to prefer running genhostid over spl.ko module parameters, and indicate that the spl.ko module needs to be reloaded after generating a new hostid.

[Cliff White (Inactive)]

The fix is simple - reload the SPL module after creating /etc/hostid. Nathaniel will be updating the documents. We are able to format now.

[Peter Jones]

Nathaniel

Can you please help improve the docs here?

Peter