[LU-3401] 2.1.5<->2.4.0 interop: sanity test 27z: FAIL: parent SEQ mismatch

• Type: Bug
• Resolution: Won't Fix
• Priority: Minor
• Fix Version/s: Lustre 2.1.6
• Affects Version/s: Lustre 2.4.0, Lustre 2.1.5
• Labels: mn1

Environment

Lustre b2_1 client build: http://build.whamcloud.com/job/lustre-b2_1/204
Lustre master server build: http://build.whamcloud.com/job/lustre-master/1508
Distro/Arch: RHEL6.4/x86_64

• Severity: 3
• Assignee: Zhenyu Xu
• Reporter: Jian Yu

Description

sanity test 27z failed as follows:

== sanity test 27z: check SEQ/OID on the MDT and OST filesystems == 06:03:20 (1369487000)
1+0 records in
1+0 records out
1048576 bytes (1.0 MB) copied, 0.0363886 s, 28.8 MB/s
7+0 records in
7+0 records out
7340032 bytes (7.3 MB) copied, 0.597308 s, 12.3 MB/s
check file /mnt/lustre/d0.sanity/d27/f27z-1
	seq 0x2000013a0, oid 0x1b19a ver 0x0
	stripe count: 1
CMD: client-26vm4 grep -c /mnt/ost1' ' /proc/mounts
Stopping /mnt/ost1 (opts:)
CMD: client-26vm4 umount -d /mnt/ost1
CMD: client-26vm4 lsmod | grep lnet > /dev/null && lctl dl | grep ' ST '
CMD: client-26vm4 mount -t ldiskfs /dev/lvm-OSS/P1 /mnt/ost1 -o loop
CMD: client-26vm4 find /mnt/ost1/O/0 -name 404
CMD: client-26vm4 /usr/sbin/ll_decode_filter_fid /mnt/ost1/O/0/d20/404
 sanity test_27z: @@@@@@ FAIL: parent SEQ mismatch

Maloo report: https://maloo.whamcloud.com/test_sets/c8747630-c621-11e2-9bf1-52540035b04c

Issue Links

• is duplicated by: LU-3393 2.1.5<->2.4.0 interop: Test timeout on test suite sanity, subtest test_31k

Comments

[James A Simmons]

Old ticket for unsupported version

[Jian Yu]

Patch for Lustre b2_1 branch to resolve the interop issue: http://review.whamcloud.com/6762

[Jian Yu]

We have to add Lustre version check codes into sanity test 27z. The failure also occurs on the combination of 2.1.6 RC2 client with 2.3.0 server:
https://maloo.whamcloud.com/test_sets/33e48fac-dd08-11e2-8c97-52540035b04c

[Zhenyu Xu]

Hi yujian,

the log shows

got: objid=290 seq=0 parent=[0x2000013a0:0x1af04:0x0]

which reveals two things: the sanity test script contains this patch, ll_decode_filter_fid does not. New ll_decode_filter_fid is supposed to print something like got: "objid=290 seq=0 parent=[0x2000013a0:0x1af04:0x0] stripe=0"

The new interop regression is due to that 2.1.5 server does not contain the ll_decode_filter_fid.c change while the sanity test27z change (which 2.1.6 contains) needs couple with it.

[Jian Yu]

would you mind trying this patch for 2.1 client http://review.whamcloud.com/6468 ?

The above patch introduced an interop regression issue:

Lustre client: v2_1_6_RC2
Lustre server: 2.1.5

sanity test 27z failed: https://maloo.whamcloud.com/test_sets/df8fce10-dcbe-11e2-b7a6-52540035b04c

[Jian Yu]

would you mind trying this patch for 2.1 client http://review.whamcloud.com/6468 ?

Hi Bobi, please add the following test parameters into the commit message so that autotest can verify the patch automatically:

Test-Parameters: envdefinitions=SLOW=yes,ENABLE_QUOTA=yes \
clientdistro=el6 serverdistro=el6 clientarch=x86_64 \
serverarch=x86_64 serverjob=lustre-master serverbuildno=1508 \
testlist=sanity

[Zhenyu Xu]

Hi Yu Jian,

would you mind trying this patch for 2.1 client http://review.whamcloud.com/6468 ?

[Peter Jones]

Bobijam

Could you please look into this one?

Thanks

Peter