Instance terminating during Backup

I'm admnistering a 350 GB Oracle 8 database. During backup datafiles ,control files & redologs are copied from one disk array to another.

Of late during backup DB writer & log writer processess are terminating the instance.

below is a part of my alert log file



Tue Sep 5 19:07:12 2006
alter tablespace MOVIMIENTOS12 begin backup
Completed: alter tablespace MOVIMIENTOS12 begin backup
Tue Sep 5 19:07:14 2006
alter tablespace IMAGENES11 begin backup
Completed: alter tablespace IMAGENES11 begin backup
Tue Sep 5 19:27:49 2006
DBW0: terminating instance due to error 27062
Instance terminated by DBW0, pid = 13928




Completed: alter tablespace INDICES77A begin backup
Mon Sep 18 19:27:08 2006
alter tablespace INDICES88 begin backup
Mon Sep 18 19:27:12 2006
Thread 1 advanced to log sequence 114798
Current log# 2 seq# 114798 mem# 0: /data/sgc/dbics/dbics06/ics.log2A
Current log# 2 seq# 114798 mem# 1: /data/sgc/dbics/admin/safety_mirror/ics.log2B
Mon Sep 18 19:27:13 2006
Completed: alter tablespace INDICES88 begin backup
Mon Sep 18 19:27:13 2006
alter tablespace MOVIMIENTOS12 begin backup
Completed: alter tablespace MOVIMIENTOS12 begin backup
Mon Sep 18 19:27:16 2006
alter tablespace IMAGENES11 begin backup
Completed: alter tablespace IMAGENES11 begin backup
Mon Sep 18 20:49:58 2006
Thread 1 advanced to log sequence 114799
Current log# 3 seq# 114799 mem# 0: /data/sgc/dbics/dbics06/ics.log3A
Current log# 3 seq# 114799 mem# 1: /data/sgc/dbics/admin/safety_mirror/ics.log3B
Mon Sep 18 21:11:28 2006
Thread 1 advanced to log sequence 114800
Current log# 1 seq# 114800 mem# 0: /data/sgc/dbics/dbics06/ics.log1A
Current log# 1 seq# 114800 mem# 1: /data/sgc/dbics/admin/safety_mirror/ics.log1B
Mon Sep 18 21:28:27 2006
Thread 1 advanced to log sequence 114801
Current log# 2 seq# 114801 mem# 0: /data/sgc/dbics/dbics06/ics.log2A
Current log# 2 seq# 114801 mem# 1: /data/sgc/dbics/admin/safety_mirror/ics.log2B
LGWR: terminating instance due to error 27062
Instance terminated by LGWR, pid = 17663



kindly advice how to deal with the problem once & for all


Regards

Msororaji

Not sure whether this will help:

http://groups.google.fr/group/comp.d...85efdf72e6985b

I saw some where in metalink (unable to find it now..!), it is an internal error because of DB_ASYNCH_IO=TRUE. It was suggesting some OS patch to resove the problem and till then convert DB_ASYNCH_IO=FALSE. Is there any associated error message in OS log?

thanks for your assistance

I checked all the OS log files & nothing was written when the instance was crashing

Regards

Found in metalink few I/O realted bugs ( specially in Sun Solaris OS) and crashing database. Did you try the backup with DB_ASYNCH_IO=FALSE? Anyway time to think about an upgarde to get support from Oracle. Raise an SR with Metalink.

Yes my OS plaftform is Solaris 5.8. Will be upgrading to Solaris 9 early next yr & later on to Oracle 9i

My init.ora file dosent have parameter DB_ASYNCH_IO. I've inserted (ie DB_ASYNCH_IO=FALSE) but since my Database is a 24X7 I guess I'll have to wait for the next crash so that the change takes effect.

Thanks

Quote:

---

Will be upgrading to Solaris 9 early next yr & later on to Oracle 9i

---

wouldn't it, in summer 2007, not make more sense to directly upgrade to 10gR2 (10.2.0.4) ?

Quote:

---

Originally Posted by Msororaji

thanks for your assistance

I checked all the OS log files & nothing was written when the instance was crashing

Regards

---

What OS logs did you look at? /var/adm/messages?

Definitely raise this one with Oracle Support.

My ITM insists I upgrade to 9i and then 10g after a year. Can't convince him otherwise. Seems to be a sceptic.

checked for logs in /var/adm/messages & for trace files in udest & bdest directories but there was non when the instance was terminating