[LU-5961] Concurrent mount of ZFS targets fails when modules are not loaded

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.7.0
• Affects Version/s: Lustre 2.5.3
• Environment: Centos6.5
• Severity: 3
• Assignee: Nathaniel Clark
• Reporter: Oliver Mangold

Description

When I mount multiple ZFS osts concurrently, e.g. by a script like the one below, I get errors like this:

—
Unable to open /dev/zfs: No such file or directory.
Verify the ZFS module stack is loaded by running '/sbin/modprobe zfs'.
Failed to initialize ZFS library
Unable to open /dev/zfs: No such file or directory.
Verify the ZFS module stack is loaded by running '/sbin/modprobe zfs'.
Failed to initialize ZFS library
Unable to open /dev/zfs: No such file or directory.
Verify the ZFS module stack is loaded by running '/sbin/modprobe zfs'.
Failed to initialize ZFS library
—

Mounting the targets serially works. Mounting in parallel also works, when the zfs and lustre modules are already loaded.

—
#!/bin/bash

mount -t lustre ost0/lnec /mnt/lustre/ost0 &
mount -t lustre ost1/lnec /mnt/lustre/ost1 &
mount -t lustre ost2/lnec /mnt/lustre/ost2 &
mount -t lustre ost3/lnec /mnt/lustre/ost3 &
mount -t lustre ost4/lnec /mnt/lustre/ost4 &
mount -t lustre ost5/lnec /mnt/lustre/ost5 &
wait
—

Syslog shows some problems about unknown symbols:
> dmesg
...
ZFS: Loaded module v0.6.3-1, ZFS pool version 5000, ZFS filesystem version 5
LNet: HW CPU cores: 24, npartitions: 4
alg: No test for crc32 (crc32-table)
alg: No test for adler32 (adler32-zlib)
alg: No test for crc32 (crc32-pclmul)
Lustre: Lustre: Build Version: 2.5.3-RC1--PRISTINE-2.6.32-431.23.3.el6_lustre.x86_64
fld: gave up waiting for init of module ptlrpc.
fld: Unknown symbol RQF_FLD_QUERY
fld: gave up waiting for init of module ptlrpc.
fld: Unknown symbol req_capsule_client_get

Issue Links

• duplicates: LU-1279 failure trying to mount two targets at the same time after boot