[Ocfs2-users] Problems with ocfs2 when rebooting the first node.

[José Costa]

Hello,

I'm using 2.6.16.41-SLES10_SP1_BRANCH_20070220135926-smp with OCFS2 1.2.4.

If I start the node1 and then the node2... everything works. If I
reboot the node1, it gives this error to node2 and I can't mount on
node1 when it comes up and can't do anything on node2 ocfs2 mounts and
also in /sys/kernel/cluster/*.

I've 8 ocfs2 partitions. (don't ask why)

Here's the kernel bug.

Feb 26 17:39:42 system2 kernel:
(3903,1):dlm_deref_lockres_handler:2353 ERROR:
5400F4D01A9E4561961EFD460CE743B9:M000000000000
0000000005e6b4c612: node 0 trying to drop ref but it is already dropped!
Feb 26 17:39:42 system2 kernel: ------------[ cut here ]------------
Feb 26 17:39:42 system2 kernel: kernel BUG at fs/ocfs2/dlm/dlmdebug.c:304!
Feb 26 17:39:42 system2 kernel: invalid opcode: 0000 [#1]
Feb 26 17:39:42 system2 kernel: SMP
Feb 26 17:39:42 system2 kernel: last sysfs file:
/devices/pci0000:00/0000:00:05.0/resource
Feb 26 17:39:42 system2 kernel: Modules linked in: ocfs2 af_packet
ocfs2_user_heartbeat ocfs2_dlmfs ocfs2_dlm ocfs2_nodemanag
er configfs bonding button battery ac apparmor aamatch_pcre loop
dm_mod i2c_piix4 i2c_core ohci_hcd sworks_agp usbcore agpgar
t e100 mii e1000 shpchp pci_hotplug ide_cd cdrom parport_pc lp parport
ext3 jbd edd fan thermal processor i2o_block i2o_core
qla2xxx firmware_class scsi_transport_fc sg st aic7xxx
scsi_transport_spi serverworks sd_mod scsi_mod ide_disk ide_core
Feb 26 17:39:42 system2 kernel: CPU:    1
Feb 26 17:39:42 system2 kernel: EIP:    0060:[<f9356fe2>]    Tainted:
G     U VLI
Feb 26 17:39:42 system2 kernel: EFLAGS: 00010202
(2.6.16.41-SLES10_SP1_BRANCH_20070220135926-smp #1)
Feb 26 17:39:42 system2 kernel: EIP is at
__dlm_print_one_lock_resource+0x12/0x729 [ocfs2_dlm]
Feb 26 17:39:42 system2 kernel: eax: f70ae401   ebx: 00000000   ecx:
00000000   edx: 00000282
Feb 26 17:39:42 system2 kernel: esi: f70ae460   edi: 0000001f   ebp:
00000001   esp: f682de54
Feb 26 17:39:42 system2 kernel: ds: 007b   es: 007b   ss: 0068
Feb 26 17:39:42 system2 kernel: Process o2net (pid: 3903,
threadinfo=f682c000 task=f4a81910)
Feb 26 17:39:42 system2 kernel: Stack: <0>00000000 00000002 f70ae460
c0130f57 f682de64 f682de64 00000005 00000082
Feb 26 17:39:42 system2 kernel:        f682de9c f65a9d64 f65a9d54
00000000 f70ae460 0000001f 00000001 c0120a80
Feb 26 17:39:42 system2 kernel:        f9374221 f682dea8 f682dea8
f936527b f9374221 00000f3f 00000001 f936e576
Feb 26 17:39:42 system2 kernel: Call Trace:
Feb 26 17:39:42 system2 kernel:  [<c0130f57>] autoremove_wake_function+0x0/0x2d
Feb 26 17:39:42 system2 kernel:  [<c0120a80>] printk+0x14/0x18
Feb 26 17:39:42 system2 kernel:  [<f936527b>]
dlm_deref_lockres_handler+0x2a6/0x3df [ocfs2_dlm]
Feb 26 17:39:42 system2 kernel:  [<f9365285>]
dlm_deref_lockres_handler+0x2b0/0x3df [ocfs2_dlm]
Feb 26 17:39:42 system2 kernel:  [<f92fc792>]
o2net_process_message+0x3e7/0x598 [ocfs2_nodemanager]
Feb 26 17:39:42 system2 kernel:  [<f92fba1d>]
o2net_recv_tcp_msg+0x55/0x60 [ocfs2_nodemanager]
Feb 26 17:39:42 system2 kernel:  [<f92fe2e6>]
o2net_rx_until_empty+0x64d/0x773 [ocfs2_nodemanager]
Feb 26 17:39:42 system2 kernel:  [<c012de26>] run_workqueue+0x78/0xb5
Feb 26 17:39:42 system2 kernel:  [<f92fdc99>]
o2net_rx_until_empty+0x0/0x773 [ocfs2_nodemanager]
Feb 26 17:39:42 system2 kernel:  [<c012e679>] worker_thread+0x0/0x10d
Feb 26 17:39:42 system2 kernel:  [<c012e755>] worker_thread+0xdc/0x10d
Feb 26 17:39:42 system2 kernel:  [<c011a53d>] default_wake_function+0x0/0xc
Feb 26 17:39:42 system2 kernel:  [<c0130e75>] kthread+0x9d/0xc9
Feb 26 17:39:42 system2 kernel:  [<c0130dd8>] kthread+0x0/0xc9
Feb 26 17:39:42 system2 kernel:  [<c0102005>] kernel_thread_helper+0x5/0xb
Feb 26 17:39:42 system2 kernel: Code: 64 d1 37 f9 0f 85 96 fe ff ff b0
01 86 05 60 d1 37 f9 89 e8 5b 5e 5f 5d c3 55 57 56 53
83 ec 60 89 44 24 08 8a 40 48 84 c0 7e 08 <0f> 0b 30 01 10 f4 36 f9 f6
05 81 de 30 f9 01 75 14 a1 84 de 30
Feb 26 17:39:43 system2 kernel:  <5>(6543,1):dlm_get_lock_resource:920
575FC4A619124A3BA677F994DF3B18F2:$RECOVERY: at least o
ne node (0) torecover before lock mastery can begin