[LU-12530] udev add/change rule loads zfs module on clients

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.14.0, Lustre 2.12.4
• Affects Version/s: Lustre 2.10.5, Lustre 2.12.2
• Severity: 3
• Assignee: Nathaniel Clark
• Reporter: Olaf Faaland

Description

Lustre loads the zfs module whenever a block device is added or changed, if Lustre was built against zfs. It unintentionally does this on clients as well as servers.

Commit 0d11a314787bc795797a016262e9bcfe86e2193e for LU-11563 added a udev rule,
lustre/conf/99-lustre-server.rules

When a block device is added or changed:
udev -> l_tunedisk
l_tunedisk -> mount_utils.c:osd_init()
osd_init() > backfs_ops[]>init() AKA zfs_init()
zfs_init() -> "/sbin/modprobe -q zfs"

This occurs if the Lustre rpms installed are a full build of Lustre, instead of just a client build.

Issue Links

• is related to: LU-10599 Possible for ID_FS_TYPE to be lustre / ldiskfs?
• is related to: LU-9551 I/O errors when lustre uses multipath devices
• is related to: LU-11736 Do not apply bulk IO tuning on MDT or MGT devices

Comments

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/36776/
Subject: LU-12530 utils: narrow l_tunedisk udev rule
Project: fs/lustre-release
Branch: b2_12
Current Patch Set:
Commit: e365cf2b001af7ca0883d293b7c148bcaa8e1c61

[Gerrit Updater]

Minh Diep ([e-mail]) uploaded a new patch: https://review.whamcloud.com/36776
Subject: LU-12530 utils: narrow l_tunedisk udev rule
Project: fs/lustre-release
Branch: b2_12
Current Patch Set: 1
Commit: f5f3f9388aa9fa2c8eb7f2e83b3eb1912afea1fd

[Peter Jones]

Landed for 2.14

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/36599/
Subject: LU-12530 utils: narrow l_tunedisk udev rule
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 7b2cb54858daa60d560fd6374c4ecba552a10d27

[Olaf Faaland]

After this lands, please consider backporting to b2_12. Thanks.

[Gerrit Updater]

Olaf Faaland-LLNL ([e-mail]) uploaded a new patch: https://review.whamcloud.com/36599
Subject: LU-12530 utils: narrow l_tunedisk udev rule
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: daa0f160640a6a3e5581e1f10ad33047a8a44ae6

[James A Simmons]

One day ext4 will be what is ldiskfs today. So Nathaniel approach is the best idea.

[Nathaniel Clark]

For ldiskfs type, I think, the closest you can get is checking the FS_LABEL, that it matches one of these:

MGS
*-MDT*
*-OST*

If it doesn't match one of those, then it definitely isn't ldiskfs.

[Olaf Faaland]

I don't see another udev property that would help distinguish ext4-formatted devices from ldiskfs-formatted devices, but if you know of one that would be even better.

[Olaf Faaland]

>> 3. zfs_tune_lustre() doesn't do anything
>
> But it might someday...

That's fair, but it shouldn't do anything with the block devices in the pool in any case. If tunables need to be set at that level, that should be handled within ZFS so that non-lustre ZFS users get the benefit as well. ZFS already has to deal properly with block devices appearing and disappearing. For example, ZFS currently sets the elevator to "noop" for the block devices in the pool.

zpool and dataset properties are persistent, so they don't need this treatment. One could also set module parameters upon module load, but this isn't the place for that.

So I think this rule should match on

ENV{ID_FS_TYPE}=="ext4"

in addition to the ACTION and SUBSYSTEM already specified. That doesn't actually solve my problem, but I think it's a step in the right direction.