[OpenAFS-devel] Re: Tracking down an oops.

[Mike Polek]

Not too much happening around the time of the oops. The box is a
firewall/gateway, so I was working on the external NIC and the
firewall rules. I do recall that AFS doesn't like NIC's going
away. Any chance that's related?

System is running a client only. I noticed that the afsd program
does have a reference to a NetInfo file. I tried dropping it in
/etc/openafs/NetInfo, but it looks like the client is still
bound to the 0.0.0.0:7001 UDP port instead of the individual
internal IP. Where does the NetInfo file go for the layout where
the config files are in /etc/openafs?


Thanks for the assistance!

--Mike

-----------  Segment of Message Log ---------------
Mar  8 11:50:02 gw1 kernel: e1000: eth1: e1000_watchdog: NIC Link is Up 100 
Mbps Full Duplex
Mar  8 11:53:09 gw1 kernel: e1000: eth1: e1000_watchdog: NIC Link is Up 100 
Mbps Full Duplex
Mar  8 11:57:10 gw1 iptables:  succeeded 
     Mar  8 12:01:01 gw1 crond(pam_unix)[29495]: session opened for user 
root by (uid
=0) 
     Mar  8 12:01:01 gw1 crond(pam_unix)[29495]: session closed for user root
Mar  8 12:10:52 gw1 kernel: afs_InitCacheInfo --- called for non-ufs 
cache!<1>Unable to handle kernel paging request at virtual address ffffffff
Mar  8 12:10:52 gw1 kernel:  printing eip: 
     Mar  8 12:10:52 gw1 kernel: f8bb929c
Mar  8 12:10:52 gw1 kernel: *pde = 00000000 
     Mar  8 12:10:52 gw1 kernel: Oops: 0002 [#1]
Mar  8 12:10:52 gw1 kernel: SMP 
     Mar  8 12:10:52 gw1 kernel: Modules linked in: libafs(U) ip_vs ipt_MARK 
iptable_
mangle ipt_mark ipt_limit ip_nat_ftp iptable_nat ip_conntrack_ftp ipt_state 
ip_conntrack ipt_REJECT ipt_LOG iptable_filter ip_tables md5 ipv6 autofs4 
i2c_dev i2
c_core sunrpc dm_mod video button battery ac uhci_hcd ehci_hcd e1000 floppy 
ext3 jbd megaraid_mbox megaraid_mm sd_mod scsi_mod
Mar  8 12:10:52 gw1 kernel: CPU:    0
Mar  8 12:10:52 gw1 kernel: EIP:    0060:[<f8bb929c>]    Tainted: P      VLI
Mar  8 12:10:52 gw1 kernel: EFLAGS: 00010206   (2.6.10-1.741_FC3smp)
Mar  8 12:10:52 gw1 kernel: EIP is at osi_Panic+0x17/0x23 [libafs]
Mar  8 12:10:52 gw1 kernel: eax: 0000002f   ebx: f8bd7f8a   ecx: cc25ce28 
edx: f8bd4ba9
Mar  8 12:10:52 gw1 kernel: esi: f6a47280   edi: f6a4729a   ebp: 0806ddc0 
esp: cc25ce24
Mar  8 12:10:52 gw1 kernel: ds: 007b   es: 007b   ss: 0068
Mar  8 12:10:52 gw1 kernel: Process afsd (pid: 29503, threadinfo=cc25c000 
task=f7a4aa40)
Mar  8 12:10:52 gw1 kernel: Stack: f8bd4ba9 00000000 00000019 00000246 
f6a47280 f8b8d17b 00000246 00000000
Mar  8 12:10:52 gw1 kernel:        00000100 f6a47280 00000246 00000019 
00000007 f6a47280 00000007 f8bc8d2a
Mar  8 12:10:52 gw1 kernel:        f7a4aa40 c1922a1c c0165697 e482d580 
cc25ce94 00000001 cc25cecc f5644cc0
Mar  8 12:10:52 gw1 kernel: Call Trace:
Mar  8 12:10:52 gw1 kernel:  [<f8b8d17b>] afs_InitCacheInfo+0x1f/0xce [libafs]
Mar  8 12:10:52 gw1 kernel:  [<f8bc8d2a>] afs_syscall_call+0xb41/0x13de [libafs]
Mar  8 12:10:52 gw1 kernel:  [<c0165697>] d_rehash+0x57/0x64
Mar  8 12:10:52 gw1 kernel:  [<c019caae>] inode_has_perm+0x4c/0x54
Mar  8 12:10:52 gw1 kernel:  [<f8bc97c1>] afs_syscall+0x148/0x3a4 [libafs]
Mar  8 12:10:52 gw1 kernel:  [<c015c413>] permission+0x43/0x48
Mar  8 12:10:52 gw1 kernel:  [<f8bc2b9b>] afs_ioctl+0x41/0x4e [libafs]
Mar  8 12:10:52 gw1 kernel:  [<c01608f4>] file_ioctl+0x188/0x19a
Mar  8 12:10:52 gw1 kernel:  [<c0160abc>] sys_ioctl+0x1b6/0x1f2
Mar  8 12:10:52 gw1 kernel:  [<c0103c97>] syscall_call+0x7/0xb
Mar  8 12:10:52 gw1 kernel: Code: e8 42 ff ff ff 89 d8 5b 5e c3 0f b7 d0 31 
c0 e9 d3 ff ff ff 53 85 c0 bb 8a 7f bd f8 ff 74 24 08 0f 44 c3 51 52 50 e8 
c1 4c 56 c7 <c6> 05 ff ff ff ff 2a 83 c4 10 5b c3 57 89 c7 83 c8 ff 56 89 d6



chas williams - CONTRACTOR wrote:
> this is probably better than any other list.  is there any chance of
> seeing oops.txt (or the chunk of the messages just before and including
> the oops)?  the 2.6 kernel usually produces a somewhat useful messages
> without ksymoops.
> 
> In message <422E5A09.6080404@pictage.com>,Mike Polek writes:
> 
>>Hi, all,
>>  This may not be the right list, but I'm hoping to get
>>a quick answer. I am using FC3, 2.6.10-1.741_FC3smp,
>>AFS version:  OpenAFS 1.3.78 built  2005-02-01
>>I got an oops in afsd, and I wanted to track it down.
>>
>>afsd -memcache -blocks 65536 -chunksize 16 -stat 2800 -daemons 5 -volumes 128