[LU-7819] sanityn test_71: @@@@@@ FAIL: data is not flushed from client

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.9.0
• Fix Version/s: Lustre 2.10.0
• Type: Bug
• Priority: Minor
• Reporter: Ashish Purkar (Inactive)
• Assignee: WC Triage
• Resolution: Fixed
• Labels: patch
• Severity: 3

Description

Steps to reproduce:

# lfs setstripe -c -1 /mnt/lustre

# lfs getstripe /mnt/lustre
/mnt/lustre
stripe_count:   -1 stripe_size:    1048576 stripe_offset:  -1

# ONLY=71 sh lustre/tests/sanityn.sh

Console Output:

== sanityn test 71: correct file map just after write operation is finished == 0
5:36:24 (1456378584)
1+0 records in
1+0 records out
40960 bytes (41 kB) copied, 0.00490768 s, 8.3 MB/s
1+0 records in
1+0 records out
40960 bytes (41 kB) copied, 0.00487972 s, 8.4 MB/s
  File: `/mnt/lustre2/f71.sanityn'
  Size: 163840          Blocks: 1          IO Block: 4194304 regular file
Device: 2c54f966h/743766374d    Inode: 144115205272502285  Links: 1
Access: (0644/-rw-r--r--)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2016-02-25 05:36:24.000000000 +0000
Modify: 2016-02-25 05:36:24.000000000 +0000
Change: 2016-02-25 05:36:24.000000000 +0000
31
fd: 3
error while ioctl 524
 sanityn test_71: @@@@@@ FAIL: data is not flushed from client
  Trace dump:
  = lustre/tests/../tests/test-framework.sh:4670:error_noexit()
  = lustre/tests/../tests/test-framework.sh:4704:error()
  = lustre/tests/sanityn.sh:2758:test_71()
  = lustre/tests/../tests/test-framework.sh:4951:run_one()
  = lustre/tests/../tests/test-framework.sh:4988:run_one_logged()
  = lustre/tests/../tests/test-framework.sh:4805:run_test()
  = lustre/tests/sanityn.sh:2775:main()
Dumping lctl log to /tmp/test_logs/1456378583/sanityn.test_71.*.1456378584.log
Dumping logs only on local client.
Resetting fail_loc on all nodes...done.
FAIL 71 (0s)
cleanup: ======================================================
== sanityn test complete, duration 1 sec == 05:36:24 (1456378584)

Issue Links

• is related to: LU-9383 ost-pools test_24: Stripe count -1 no...

Comments

[Ashish Purkar (Inactive)]

• With stripecount = 1 (default), sanityn test_71 passed on master.
• With stripecount = -1, it failed with ioctl error code 524 i.e. LUSTRE_ENOTSUPP
• Respective code area:

  static const struct cl_object_operations lov_ops = {
          .coo_fiemap       = lov_object_fiemap,
  ...		
  static int lov_object_fiemap(const struct lu_env *env, struct cl_object *obj,
                               struct ll_fiemap_info_key *fmkey,
                               struct fiemap *fiemap, size_t *buflen)
  -->
          /**
           * If the stripe_count > 1 and the application does not understand
           * DEVICE_ORDER flag, it cannot interpret the extents correctly.
           */
          if (lsm->lsm_stripe_count > 1 && !(fiemap->fm_flags &
                                             FIEMAP_FLAG_DEVICE_ORDER))
                  GOTO(out_lsm, rc = -ENOTSUPP);
  

• root cause of the problem in checkfiemap test tool where "fiemap->fm_flags = FIEMAP_FLAG_SYNC" set only and FIEMAP_FLAG_DEVICE_ORDER is not set.
• Solution here is to add FIEMAP_FLAG_DEVICE_ORDER in flag which is supported by lustre fiemap implementation.

[Gerrit Updater]

Ashish Purkar ([e-mail]) uploaded a new patch: http://review.whamcloud.com/18679
Subject: LU-7819 tests: fix checkfiemap for stripecount > 1
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 99c441c7b62afff48ce8a6ed02804a70e390b6af

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/18679/
Subject: LU-7819 tests: fix checkfiemap for stripecount > 1
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 8b8bc6eee05758cec346f1f9aeed92f3334d8cf3

[Peter Jones]

Landed for 2.10