Re: [suse-oracle] ocfs2 problem after sles 10 SP2 upgrade

Hello Draghici,

#1 & #3 are caused by same issue, and most likely addressed in latest SP2 kernel update (2.6.16.60-0.37).

#2 - It appears to be OCFS2 related. 

Please apply latest kernel patch and If it doesn't help open a bug with Novell.

Thanks,
Arun

>>> On 4/13/2009 at 6:53 AM, gabi draghici <[email protected]> wrote:
> Hello to everyone !
> 
> We had an sles 10 Sp1 , oracle rac 10.2.0.4,  ocfs  combination wich worked
> fine for about an year !
> After we've done some testing we decided to upgrade to sles 10  Sp2 !
> After upgrade we've notice 3 problems so far :
> 
>    1. rman ( oracle tool for backup) doesn't work anymore from the local fs
> :
> 
>  Recovery Manager: Release 10.2.0.4.0 - Production on Mon Apr 13 16:43:19
> 2009
>  Copyright (c) 1982, 2007, Oracle.  All rights reserved.
>  RMAN-00571: ===========================================================
>  RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
>  RMAN-00571: ===========================================================
>  RMAN-00554: initialization of internal recovery manager package failed
>  RMAN-03000: recovery manager compiler component initialization failed
>  RMAN-06001: error parsing job step library
>  RMAN-01006: error signalled during parse
>  RMAN-00600: internal error, arguments [8083] [] [] [] []
>  LFI-00005: Free some memory failed in lfibrdt().
>  LFI-00004: Call to lfibgl() failed.
> 
>    2. in /var/log/messages we have :
> 
> 
> Apr 13 13:59:03 nod2 kernel: Call Trace:
> <ffffffff8016357a>{remove_from_page_cache+49}
> Apr 13 13:59:03 nod2 kernel:
> <ffffffff801695b4>{truncate_complete_page+53}
> <ffffffff8016965e>{truncate_inode_pages_range+159}
> Apr 13 13:59:03 nod2 kernel:
> <ffffffff884af364>{:ocfs2:ocfs2_data_convert_worker+202}
> Apr 13 13:59:03 nod2 kernel:
> <ffffffff884ad602>{:ocfs2:ocfs2_downconvert_thread+1190}
> Apr 13 13:59:03 nod2 kernel:
> <ffffffff80148092>{autoremove_wake_function+0}
> <ffffffff884ad15c>{:ocfs2:ocfs2_downconvert_thread+0}
> Apr 13 13:59:03 nod2 kernel:
> <ffffffff80147c88>{keventd_create_kthread+0} <ffffffff80147f50>{kthread+236}
> Apr 13 13:59:04 nod2 kernel:        <ffffffff8010bed2>{child_rip+8}
> <ffffffff80147c88>{keventd_create_kthread+0}
> Apr 13 13:59:04 nod2 kernel:        <ffffffff80147e64>{kthread+0}
> <ffffffff8010beca>{child_rip+0}
> Apr 13 13:59:04 nod2 kernel: Badness in __remove_from_page_cache_nocheck at
> mm/filemap.c:122
> 
> The only thing we found about that is a novell document, id = 7000562 wich
> resolution is "please open a service request ... ".
> 
> 
>  3. in oracle's alert.log we found that arch process can't write :
> 
> 
> 
> ARC1: Encountered disk I/O error 19502
> Mon Apr 13 13:02:36 2009
> ARC1: Closing local archive destination LOG_ARCHIVE_DEST_1:
> '/oracle/PRD/oraarch/PRDarch/1_12508_639435084.dbf' (error 19502)
>  (PRD001)
> Mon Apr 13 13:02:36 2009
> Errors in file /oracle/PRD/saptrace/background/prd001_arc1_9381.trc:
> ORA-19502: write error on file
> "/oracle/PRD/oraarch/PRDarch/1_12508_639435084.dbf", blockno 32769
> (blocksize=512)
> ORA-27061: waiting for async I/Os failed
> Linux-x86_64 Error: 5: Input/output error
> Additional information: -1
> Additional information: 1048576
> 
> 
> This one we solved by disabling the disk_asynch_io and filesystemio_options
> ! The efect is that performance is affected (but at least the arch
> process is ok now );
> 
> 
> Any help is appreciated !
> 
> 
> Draghici Gabriel
> database administrator

_______________________________________________
suse-oracle mailing list
[email protected]
http://listx.novell.com/mailman/listinfo/suse-oracle