Re: IO errors on Linux x86_64

On 6/8/06, Kevin Closson <kevinc@xxxxxxxxxxxxx> wrote: 

        Storage:  Dell Clariion CX700 SAN

...powerpath ? 


Dunno yet, but I think so.


Errors in file 
/u01/app/oracle//admin/pr09/bdump/pr09_ckpt_29193.trc:
        ORA-01171: datafile 12 going offline due to error advancing
checkpoint
        ORA-01110: data file 12:
'/u01/oradata/pr09/agile_indx101pr09.ora'
                                 ^^^^^^^^^^^^^
...                              Is this ext3 ?




Yes, this is ext3.


ORA-01115: IO error reading block from file 12 (block # 1) 
        ORA-27072: skgfdisp: I/O error
        Linux-x86_64 Error: 5: Input/output error

... are there corresponding /var/log/messages entries for these EIO
errors tht would impugn the scsi layer or other goodies downwind? Trying
to establish if this is an EIO pushed up from a software layer (perhaps
powerpath) ?   Being intermittant there whould be /var/log/messages
goodies to clue us in




No IO error messages that I can find yet, but did find something else:

Jun  7 23:15:07 ordb02 kernel: ctxhx[3472]: segfault at 00000000000007b5 rip
00000000f7d240f6 rsp 00000000ffffa528 error 4
Jun  7 23:17:53 ordb02 kernel: ctxhx[3972]: segfault at 00000000000007b5 rip
00000000f7d240f6 rsp 00000000ffffa528 error 4
Jun  7 23:31:39 ordb02 kernel: ctxhx[6551]: segfault at 00000000f790c02c rip
00000000f7ea39c3 rsp 00000000ffffa370 error 6
Jun  8 09:55:05 ordb02 kernel: ctxhx[18246]: segfault at 00000000000007b5
rip 00000000f7d240f6 rsp 00000000ffffb1b8 error 4

There are 2 other databases running the same app, same versions of Oracle,
so it
is a little difficult to be sure which is causing the problem.

Since most of these are the past 2 days, it could be that this is just a
symptom.  The IO errors
may have crashed ctxhx (I'm assuming it has something to do with building
Text indexes)

--
Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist

Other related posts: