[LU-9510] Kernel panic - not syncing: LDISKFS-fs (device md1): panic forced after error: LDISKFS-fs error (device md1): ldiskfs_mb_release_inode_pa: pa free mismatch: [pa ffff8807c80eb828] [phy 32256] [logic 20] [len 20] [free 19] [error 0] [inode 80] [freed 20]

• Status: Resolved
• Project: Lustre
• Fix Version/s: Lustre 2.15.0
• Type: Bug
• Priority: Minor
• Reporter: VIKRAM BABASO JADHAV (Inactive)
• Assignee: Yang Sheng
• Resolution: Fixed
• Labels: patch
• Severity: 3

Description

2016-08-02 12:12:50 [319476.675751] LustreError: 55339:0:(import.c:671:ptlrpc_connect_import()) can't connect to a closed import
2016-08-02 12:12:50 [319476.685482] LustreError: 55339:0:(import.c:671:ptlrpc_connect_import()) Skipped 1 previous similar message
2016-08-02 12:12:51 [319476.844638] LDISKFS-fs error (device md1): ldiskfs_mb_release_inode_pa: pa free mismatch: [pa ffff8807c80eb828] [phy 32256] [logic 20] [len 20] [free 19] [error 0] [inode 80] [freed 20]
2016-08-02 12:12:51 [319476.861543] Aborting journal on device md1-8.
2016-08-02 12:12:51 [319476.883283] Kernel panic - not syncing: LDISKFS-fs (device md1): panic forced after error
2016-08-02 12:12:51 [319476.883286]
2016-08-02 12:12:51 [319476.893384] Pid: 93703, comm: umount Tainted: P --------------- 2.6.32-431.17.1.x2.1.32.x86_64 #1
2016-08-02 12:12:51 [319476.903624] Call Trace:
2016-08-02 12:12:51 [319476.906249] [<ffffffff8152562e>] ? panic+0xa7/0x16f
2016-08-02 12:12:51 [319476.911414] [<ffffffffa0f41998>] ? ldiskfs_commit_super+0x188/0x210 [ldiskfs]
2016-08-02 12:12:51 [319476.918884] [<ffffffffa0f41fa4>] ? ldiskfs_handle_error+0xc4/0xd0 [ldiskfs]
2016-08-02 12:12:51 [319476.926111] [<ffffffffa0f42342>] ? __ldiskfs_error+0x82/0x90 [ldiskfs]
2016-08-02 12:12:51 [319476.932904] [<ffffffffa0f26630>] ? ldiskfs_mb_release_inode_pa+0x250/0x3a0 [ldiskfs]
2016-08-02 12:12:51 [319476.940977] [<ffffffff811bf706>] ? __wait_on_buffer+0x26/0x30
2016-08-02 12:12:51 [319476.946988] [<ffffffffa0f2ae1e>] ? ldiskfs_discard_preallocations+0x1fe/0x490 [ldiskfs]
2016-08-02 12:12:51 [319476.955325] [<ffffffffa0f3f9e6>] ? ldiskfs_clear_inode+0x16/0x50 [ldiskfs]
2016-08-02 12:12:51 [319476.962459] [<ffffffff811a614c>] ? clear_inode+0xac/0x140
2016-08-02 12:12:51 [319476.968116] [<ffffffff811a6220>] ? dispose_list+0x40/0x120
2016-08-02 12:12:51 [319476.973855] [<ffffffff811a66ca>] ? invalidate_inodes+0xea/0x190
2016-08-02 12:12:51 [319476.980040] [<ffffffff8118b77c>] ? generic_shutdown_super+0x4c/0xe0
2016-08-02 12:12:51 [319476.986565] [<ffffffff8118b841>] ? kill_block_super+0x31/0x50
2016-08-02 12:12:51 [319476.992570] [<ffffffff8118c017>] ? deactivate_super+0x57/0x80
2016-08-02 12:12:51 [319476.998580] [<ffffffff811ab01f>] ? mntput_no_expire+0xbf/0x110
2016-08-02 12:12:51 [319477.004681] [<ffffffffa0f92b8d>] ? osd_umount+0x5d/0x130 [osd_ldiskfs]
2016-08-02 12:12:51 [319477.011485] [<ffffffffa0f9bde7>] ? osd_device_fini+0x147/0x190 [osd_ldiskfs]
2016-08-02 12:12:51 [319477.018908] [<ffffffffa06a3252>] ? class_cleanup+0x562/0xd20 [obdclass]
2016-08-02 12:12:51 [319477.025806] [<ffffffffa06a4f7a>] ? class_process_config+0x156a/0x1ad0 [obdclass]
2016-08-02 12:12:51 [319477.033559] [<ffffffffa069de75>] ? lustre_cfg_new+0x4f5/0x6f0 [obdclass]
2016-08-02 12:12:51 [319477.040542] [<ffffffffa06a5659>] ? class_manual_cleanup+0x179/0x700 [obdclass]
2016-08-02 12:12:51 [319477.048142] [<ffffffffa0682fa9>] ? class_export_put+0x189/0x2e0 [obdclass]
2016-08-02 12:12:51 [319477.055280] [<ffffffffa0f9d6c5>] ? osd_obd_disconnect+0x1c5/0x1d0 [osd_ldiskfs]
2016-08-02 12:12:51 [319477.062943] [<ffffffffa06a7c3b>] ? lustre_put_lsi+0x1ab/0x11a0 [obdclass]
2016-08-02 12:12:51 [319477.070019] [<ffffffffa06b0458>] ? lustre_common_put_super+0x5d8/0xbf0 [obdclass]
2016-08-02 12:12:51 [319477.077858] [<ffffffffa06e2e85>] ? server_put_super+0x1d5/0xfd0 [obdclass]
2016-08-02 12:12:51 [319477.084989] [<ffffffff8118b78b>] ? generic_shutdown_super+0x5b/0xe0
2016-08-02 12:12:51 [319477.091516] [<ffffffff8118b876>] ? kill_anon_super+0x16/0x60
2016-08-02 12:12:51 [319477.097459] [<ffffffffa06a7546>] ? lustre_kill_super+0x36/0x60 [obdclass]
2016-08-02 12:12:51 [319477.104510] [<ffffffff8118c017>] ? deactivate_super+0x57/0x80
2016-08-02 12:12:51 [319477.110513] [<ffffffff811ab01f>] ? mntput_no_expire+0xbf/0x110
2016-08-02 12:12:51 [319477.116609] [<ffffffff811abb6b>] ? sys_umount+0x7b/0x3a0
2016-08-02 12:12:51 [319477.122181] [<ffffffff8108a391>] ? sigprocmask+0x71/0x110
2016-08-02 12:12:51 [319477.127843] [<ffffffff8100b072>] ? system_call_fastpath+0x16/0x1b

Comments

[Alex Zhuravlev]

was read-only (in any form) enabled on the device prior to the failure?

[Gerrit Updater]

jadhav.vikram ([e-mail]) uploaded a new patch: https://review.whamcloud.com/27130
Subject: LU-9510 ldiskfs: to not verify preallocation in umount path
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 350e368e706076d0902e6f3e60419c8519e96882

[VIKRAM BABASO JADHAV (Inactive)]

> was read-only (in any form) enabled on the device prior to the failure?

Here i am not sure about read only 

[Peter Jones]

Yang Sheng

Can you check whether this is fixed upstream already and, if not, help see that it gets the necessary reviews?

Thanks

 

Peter

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/27130/
Subject: LU-9510 ldiskfs: to not verify preallocation in umount path
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 9059a1de8913a6e81ada3f20080af0e18e4b498e

[Yang Sheng]

Patch landed. Close ticket.