I have a test environment for Oracle single instance on ASM

I have restored database from tape and during the recovery, the system crashed by some I/O error.

I restarted the server, mounted the database and run validate on archivelog(thinking that the blocks in this archivelog may be corrupted)

RMAN> validate archivelog sequence=2994;

RMAN-03009: failure of validate command on ORA_DISK_1 channel at 02/25/2014 12:24:13
ORA-19501: read error on file "+DATA01/crimesys/archivelog/2014_02_24/thread_1_seq_2994.294.8
ORA-15081: failed to submit an I/O operation to a disk
ORA-15081: failed to submit an I/O operation to a disk

And the system crushed again.

Note: Other validates are successful just this archivelog causes the system crash.


--Asm disk

ls -la /dev/oracleasm/asm_disk1
brw-rw---- 1 grid asmdba 253, 2 Feb 25 13:13 /dev/oracleasm/asm_disk1

--Trace info

more /u01/app/oracle/diag/rdbms/crimesys/crimesys/trace/crimesys_ora_2696.

WARNING: Read Failed. group:1 disk:0 AU:114682 offset:512 size:1048064
         incarnation:0xe968a075 asynchronous result:'I/O error'
         subsys:System iop:0x7fd93c64b680 bufp:0x7fd93b64e000 osderr:0x0 osderr1:0x0
WARNING: failed to read mirror side 1 of virtual extent 13 logical extent 0 of file 294 in gr
oup [1.353915012] from disk DATA01_0000  allocation unit 114682 reason error; if possible, wi
ll try another mirror side

--Also when I check the disk of ASM it is successful.

SQL> alter diskgroup data01 check 

NOTE: starting check of diskgroup DATA01
Tue Feb 25 13:36:59 2014
GMON checking disk 0 for group 1 at 5 for pid 18, osid 2181
SUCCESS: check of diskgroup DATA01 found no errors
SUCCESS: alter diskgroup data01 check

How to solve? Is the virtual block corrupted? How to check? Please help.

  • Disk permissions? 660 looks like it might cause problems. See deryaoktay.wordpress.com/tag/ora-15081 – Colin 't Hart Feb 25 '14 at 9:55
  • @Colin'tHart I have changed permissions but still not working. Maybe there is the problem with virtual disk, but don't know how to check. – kupa Feb 25 '14 at 13:33

If your are using multi-path make sure that your O.S. does not point to the file system instead of the disk groups. it happened to me 2 times after box reboot.

This is to answer your question:

SQL> set pagesize 800
SQL> select name, label, path from v$asm_disk;
NAME                           LABEL
------------------------------ -------------------------------

DGER_0000    /dev/oracleasm/disks/DSK5

DGSTBY_0000 /dev/oracleasm/disks/DSKPATHY

run the follwing command :

 /etc/init.d/oracleasm querydisk -d <your disk here>

post the output !

  • It is not using multi-path. How to check what you said? ("O.S. does not point to the file system instead of the disk groups") – kupa Feb 25 '14 at 13:38
  • It returns DATA01_0000 | /dev/oracleasm/asm_disk1 – kupa Feb 25 '14 at 14:13
  • -bash: /etc/init.d/oracleasm: No such file or directory – kupa Feb 25 '14 at 14:30
  • I am using UDEV rules not asmlib, as I remember /etc/init.d/oracleasm exists when using asmlib,isn't it? – kupa Feb 25 '14 at 14:31
  • Not sure how to deal with it now !! but take a look at this link it might help you start your troubleshhoting oracle-base.com/articles/linux/… – Up_One Feb 25 '14 at 14:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.