Summary: | openafs-1.4.1-r1 - openafs-client doesn't start | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Sven Müller <musv> |
Component: | Current packages | Assignee: | Stefaan De Roeck (RETIRED) <stefaan> |
Status: | RESOLVED INVALID | ||
Severity: | normal | CC: | net-fs |
Priority: | High | ||
Version: | 2006.0 | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Sven Müller
2006-07-02 04:03:50 UTC
I'm not sure where this error comes from, I've never experienced it myself with this version of openafs. I'm hoping the fixes introduced in openafs-1.4.2-beta solve your problem, and am planning to put up an ebuild for you to test if that's ok. Yes of course. Where can I download the ebuild? Please take a look at http://dev.gentoo.org/~stefaan/prerelease/openafs/ Thanks for your help, but it didn't solute the problem. The new (and seems to be the old) issue is:
> /etc/init.d/openafs-client start
* Starting OpenAFS client ...
* Loading OpenAFS kernel module ... [ ok ]
* Starting OpenAFS daemon ...
Oops: 0002 [#1]
PREEMPT
Modules linked in: openafs vmnet vmmon
CPU: 0
EIP: 0060:[<fa547f5e>] Tainted: P VLI
EFLAGS: 00010286 (2.6.17-gentoo-r1 #1)
EIP is at osi_Panic+0x23/0x2b [openafs]
eax: 00000016 ebx: f05fa000 ecx: ffffffff edx: 00000001
esi: ee097dac edi: fffffffb ebp: 00000000 esp: ed54cdc0
ds: 007b es: 007b ss: 0068
Process afsd (pid: 335, threadinfo=ed54c000 task=ec831530)
Stack: fa550512 fa563f0f 00000000 00000016 00000000 fa5b0760 00000001 fa51a0a2
00000000 000007ff ec831530 c03271cb ed54cdf4 22222222 22222222 22222222
0000000a ed54c000 ed54cef4 00000003 fa557452 00000000 00000000 00015895
Call Trace:
<fa550512> osi_UFSOpen+0x138/0x194 [openafs] <fa51a0a2> afs_InitCacheFile+0x201/0x417 [openafs]
<c03271cb> __mutex_lock_slowpath+0x2c8/0x37d <fa557452> afs_syscall_call+0x901/0x1342 [openafs]
<c01e6d7b> _atomic_dec_and_lock+0x27/0x44 <c0165675> mntput_no_expire+0x14/0x90
<c015aefd> link_path_walk+0xb0/0xbb <c01e6d7b> _atomic_dec_and_lock+0x27/0x44
<c0165675> mntput_no_expire+0x14/0x90 <fa5584e8> afs_syscall+0x108/0x1562 [openafs]
<c0159d26> vfs_permission+0x13/0x17 <c0159d88> may_open+0x5e/0x1a7
<c015be83> open_namei+0x248/0x591 <c014b8c5> __dentry_open+0x11d/0x204
<fa551036> afs_unlocked_ioctl+0x46/0x51 [openafs] <c015cdb5> do_ioctl+0x21/0x5f
<c015d04f> vfs_ioctl+0x25c/0x26f <c0159f13> putname+0x27/0x2a
<c015d08f> sys_ioctl+0x2d/0x46 <c01026d3> sysenter_past_esp+0x54/0x75
Code: 00 66 89 46 7c 5b 5e c3 8b 54 24 04 b8 b7 33 56 fa ff 74 24 10 ff 74 24 10 ff 74 24 10 85 d2 0f 44 d0 52 e8 76 d0 bc c5 83 c4 10 <c6> 05 00 00 00 00 00 c3 56 53 8
b 74 24 10 0f bf c6 50 ff 74 24
EIP: [<fa547f5e>] osi_Panic+0x23/0x2b [openafs] SS:ESP 0068:ed54cdc0
I don't know if it's only a bug in openafs. Maybe there's a misconfiguration in my system. But I don't have any idea, what it could be.
In this trail I installed a new kernel and rebooted the machine, so there were no other modules loaded (apart the vmware modules). Could it depend of any driver what is needed as module? Because I built nearly every possible driver directly into the kernel and not as module. So if you have any ideas for testing other kernel configurations, please tell me.
I found the error. In /etc/openafs/cacheinfo I spezified the cache-size: /afs:/var/cache/openafs:185000 and it had worked well for months. Because I'm using reiserfs for file system. I created a 200mb file and formatted this with ext2. This file I mounted to /var/cache/openafs. It worked without problems. But one month ago I had a harddisc-crash and I had to save all the data to another harddisc. And I guess there occured an error with this file, because the file size of my afs-cache-file was only 120 mb. So now I tried starting the afsd directly and received a message of an invalid inode at my cache-file. I just created a new file and now it works perfectly - version 1.4.2-beta1 too. So, have a lot thanks for your help and I'm very sorry for bugging you with this error that was my fault. But I couldn't guess nothing about the true error in the posted error message. No problem. I'm glad you found the problem. Reopening to close as invalid (else this report would stop openafs from going to stable) Problem caused by filesystem error, thus closing as invalid. (In reply to comment #5) > I found the error. > In /etc/openafs/cacheinfo I spezified the cache-size: > /afs:/var/cache/openafs:185000 > and it had worked well for months. > > Because I'm using reiserfs for file system. I created a 200mb file and > formatted this with ext2. This file I mounted to /var/cache/openafs. It worked > without problems. Just in case other run ito the same problem: OpenAFS devs tell us since years: You can't use anything else than ext2 for the cache. For /vicepXX, use whatever fs you want. |