[Gluster-devel] servers keep hanging with fuse errors

Jordi Moles jordi at cdmon.com
Wed Jan 30 10:31:14 UTC 2008 

Hi,

yes, everything is up to date, i made a new install, with a brand new 
etch virtual machine and then i used a development repository to install 
gluster with "apt-get install"

these are all the packages i've got related to either gluster or fuse:

ii  fuse-utils             2.7.2-glfs8                          
Filesystem in USErspace (utilities)
ii  glusterfs-client       1.3.8-TLA643                         
GlusterFS fuse client
ii  libfuse2               2.7.2-glfs8                          
Filesystem in USErspace library
ii  libglusterfs0          1.3.8-TLA643                         
GlusterFS libraries and translator modules

Do you think i should install form the source code and ./configure 
instead of apt-get install?
May be i can recompile now? I don't know if from the repository this 
"--enable-kernel-module" thing is set or not

En/na Anand Avati ha escrit:
> Jordi,
>  is the fuse kernel module from 2.7.2glfs8 as well? (compiled with
> ./configure --enable-kernel-module and make install inside kernel/ subdir)
> the kernel panic you've pasted was a well known issue with older fuse kernel
> modules and it seems to me that your kernel module is still old.
>
> avati
>
> 2008/1/30, Jordi Moles <jordi at cdmon.com>:
>   
>> Hi,
>>
>> i'm running this versions of the packages now:
>>
>> ii  fuse-utils
>> 2.7.2-glfs8                              Filesystem in USErspace
>> (utilities)
>> ii  glusterfs-client
>> 1.3.8-TLA643                             GlusterFS fuse client
>> ii  libfuse2
>> 2.7.2-glfs8                              Filesystem in USErspace library
>>
>> I also removed any kind of lock method from dovecot.conf and added the
>> lock function to the filesystem itself, on the server side, with the
>> "type features/posix-locks" option.
>> The thing is that now, dovecots work grate, but postfixs hang very
>> often, with the following error:
>>
>> ***************
>>
>> Unable to handle kernel paging request at 0000000000100108 RIP:
>> [<ffffffff88020838>] :fuse:request_end+0x45/0x109
>> PGD 1f29c067 PUD 1f327067 PMD 0
>> Oops: 0002 [1] SMP
>> CPU 0
>> Modules linked in: ipv6 fuse dm_snapshot dm_mirror dm_mod
>> Pid: 723, comm: glusterfs Not tainted 2.6.18-xen #1
>> RIP: e030:[<ffffffff88020838>]  [<ffffffff88020838>]
>> :fuse:request_end+0x45/0x109
>> RSP: e02b:ffff88001ecb1d68  EFLAGS: 00010246
>> RAX: 0000000000200200 RBX: ffff88001e82af48 RCX: ffff88001e82af58
>> RDX: 0000000000100100 RSI: ffff88001e82af48 RDI: ffff88001f68d400
>> RBP: ffff88001f68d400 R08: 000000001f74ab40 R09: ffff88001e82b048
>> R10: 0000000000000008 R11: ffff88001ecb1cf0 R12: 0000000000000000
>> R13: ffff88001e82af80 R14: ffff88001ecb1df8 R15: 0000000000000001
>> FS:  00002ba3e961aae0(0063) GS:ffffffff804cd000(0000)
>> knlGS:0000000000000000
>> CS:  e033 DS: 0000 ES: 0000
>> Process glusterfs (pid: 723, threadinfo ffff88001ecb0000, task
>> ffff88001fe80820)
>> Stack:  ffff88001e82af48 ffff88001f68d400 00000000fffffffe
>> ffffffff88021056
>> ffff88001ecb1ef8 000000301fcb8180 00000156fffffff4 ffff88001f72f100
>> 0000000000000015 0000000000000000 ffff88001ecb1e18 0000000041000a90
>> Call Trace:
>> [<ffffffff88021056>] :fuse:fuse_dev_readv+0x385/0x435
>> [<ffffffff802801d3>] do_readv_writev+0x271/0x294
>> [<ffffffff802274c7>] default_wake_function+0x0/0xe
>> [<ffffffff88021120>] :fuse:fuse_dev_read+0x1a/0x1f
>> [<ffffffff802804bc>] vfs_read+0xcb/0x171
>> [<ffffffff8028089b>] sys_read+0x45/0x6e
>> [<ffffffff8020a436>] system_call+0x86/0x8b
>> [<ffffffff8020a3b0>] system_call+0x0/0x8b
>>
>>
>> Code: 48 89 42 08 48 89 10 48 c7 41 08 00 02 20 00 f6 46 30 08 48
>> RIP  [<ffffffff88020838>] :fuse:request_end+0x45/0x109
>> RSP <ffff88001ecb1d68>
>> CR2: 0000000000100108
>>   postfix01gluster01 kernel: Oops: 0002 [1] SMP
>>                                                  postfix01gluster01
>> kernel: CR2: 0000000000100108
>>
>> <3>BUG: soft lockup detected on CPU#0!
>>
>> Call Trace:
>> <IRQ> [<ffffffff80257f78>] softlockup_tick+0xd8/0xea
>> [<ffffffff8020f110>] timer_interrupt+0x3a9/0x405
>> [<ffffffff80258264>] handle_IRQ_event+0x4e/0x96
>> [<ffffffff80258350>] __do_IRQ+0xa4/0x105
>> [<ffffffff8020b0e8>] call_softirq+0x1c/0x28
>> [<ffffffff8020cecb>] do_IRQ+0x65/0x73
>> [<ffffffff8034a8c1>] evtchn_do_upcall+0xac/0x12d
>> [<ffffffff8020ac1e>] do_hypervisor_callback+0x1e/0x2c
>> <EOI> [<ffffffff803f1234>] .text.lock.spinlock+0x2/0x8a
>> [<ffffffff88020a4f>] :fuse:fuse_dev_writev+0xb8/0x31b
>> [<ffffffff88020cb2>] :fuse:fuse_dev_write+0x0/0x1f
>> [<ffffffff802800d7>] do_readv_writev+0x175/0x294
>> [<ffffffff88020cb2>] :fuse:fuse_dev_write+0x0/0x1f
>> [<ffffffff803efb3b>] schedule_timeout+0x1e/0xad
>> [<ffffffff803f0976>] __down_read+0x12/0xec
>> [<ffffffff80280695>] sys_writev+0x45/0x93
>> [<ffffffff8020a436>] system_call+0x86/0x8b
>> [<ffffffff8020a3b0>] system_call+0x0/0x8b
>>
>>
>> ***************
>>
>> i set the log level for glusterfs on both nodes and clients to WARNING.
>> However, nothing was written on any of them :( .
>>
>> I keep using virtual machines with xen 3.1 to test all this, but i
>> already tried with non-virtual environments and got the same errors.
>>
>> Do you have any ideas?
>>
>> En/na Jordi Moles Blanco ha escrit:
>>     
>>> hi,
>>>
>>> thanks for the details, i'll give it a try and come back to you to tell
>>> whether has become stable or not.
>>>
>>> Thank you very much.
>>>
>>>
>>> El dt 29 de 01 del 2008 a les 01:11 +0530, en/na Anand Avati va
>>> escriure:
>>>
>>>       
>>>> Jordi,
>>>>  you should really upgrade your fuse kernel module which will fix this
>>>> issue. please use the kernel module from -
>>>>
>>>> http://ftp.zresearch.com/pub/gluster/glusterfs/fuse/
>>>>
>>>> thanks,
>>>> avati
>>>>
>>>> 2008/1/28, Jordi Moles <jordi at cdmon.com>:
>>>>
>>>>         Hi,
>>>>
>>>>         i'm sorry but i can't get any newer version with the
>>>>         repositories you
>>>>         gave me, after apt-update, apt-upgrade says there's nothing to
>>>>         upgrade.
>>>>         And i try, for example, to get the source code of every
>>>>         package, i get this
>>>>
>>>>         fuse_2.5.3-4.4, which is even older than the one installed on
>>>>         my system.
>>>>
>>>>
>>>>
>>>>
>>>>         
>>>
>>> _______________________________________________
>>> Gluster-devel mailing list
>>> Gluster-devel at nongnu.org
>>> http://lists.nongnu.org/mailman/listinfo/gluster-devel
>>>
>>>       
>>
>> _______________________________________________
>> Gluster-devel mailing list
>> Gluster-devel at nongnu.org
>> http://lists.nongnu.org/mailman/listinfo/gluster-devel
>>
>>     
>
>
>
>

More information about the Gluster-devel mailing list