a bug in 2.6.24 resulting the client "breaking" ONTAP?

[Timo Aaltonen]

On Sun, 25 May 2008, J. Bruce Fields wrote:

> On Fri, May 23, 2008 at 05:50:22PM +0300, Timo Aaltonen wrote:
>>
>>  	Hi
>>
>>    We are seeing a strange bug which makes all the machines using
>> nfs4+krb5i mounts getting "remote io error" notices when trying to access
>> the mounts.
>
> Is this an intermittent thing, or does it always happen immediately on
> first access of the mount?

It takes from one day to couple of weeks to bump into this, so it's hard 
to debug..

> Is there anything in the system logs?

Well, nothing except this crash some time before the problems started, 
but that doesn't mean they are related:

kernel BUG at /build/buildd/linux-2.6.24/net/sunrpc/rpcb_clnt.c:322!
<snip>
Call Trace:
  [sunrpc:queue_delayed_work+0x51/0x70] queue_delayed_work+0x51/0x70
  [lock_timer_base+0x27/0x60] lock_timer_base+0x27/0x60
  [<f8e65195>] call_bind+0x45/0x80 [sunrpc]
  [shpchp:del_timer_sync+0xe/0x20] del_timer_sync+0xe/0x20
  [<f8e6c07c>] rpc_delete_timer+0x2c/0x50 [sunrpc]
  [<f8e6c542>] __rpc_execute+0x62/0x280 [sunrpc]
  [10295.051592] [<f8e660c2>] rpc_release_client+0x32/0x60 [sunrpc]
  [<f8e6c760>] rpc_async_schedule+0x0/0x10 [sunrpc]
  [run_workqueue+0xbf/0x160] run_workqueue+0xbf/0x160
  [worker_thread+0x0/0xe0] worker_thread+0x0/0xe0
  [worker_thread+0x84/0xe0] worker_thread+0x84/0xe0
  [<c0140b70>] autoremove_wake_function+0x0/0x40
  [worker_thread+0x0/0xe0] worker_thread+0x0/0xe0
  [kthread+0x42/0x70] kthread+0x42/0x70
  [kthread+0x0/0x70] kthread+0x0/0x70
  [kernel_thread_helper+0x7/0x10] kernel_thread_helper+0x7/0x10

  =======================
Code: ac 00 00 00 c7 44 24 08 04 9a e7 f8 c7 04 24 30 d1 e7 f8 89 44 24 04 
e8 2f 89 2b c7 c7 44 24 20 a0 ff ff ff e9 0c fe ff ff 66 90 <0f> 0b eb fe 
0f b7 b0 ac 00 00 00 8b 57 24 8b 4f 20 8b 5f 2c 8b
  EIP: [<f8e74540>] rpcb_getport_async+0x290/0x430 [sunrpc] SS:ESP 
0068:dfb99f00
  ---[ end trace d8ca4a218e8bc80c ]---

t