IO Error

[dba_essex]

This is the message in our alert log.
We've had this error in the past and database had to be recovered. This time I haven't had to recover db - Database is up.

ORA-01110: data file 5: 'U:\ORACLE\ORADATA\EVOLVLIV\CL_IND01.DBF'
ORA-01114: IO error writing block to file 5 (block # 1)
ORA-27072: skgfdisp: I/O error
OSD-04008: WriteFile() failure, unable to write to file
O/S-Error: (OS 5) Access is denied.

[jmodic]

Is this partition U: from a local disk, or is it mapped drive from some other server?

If it is a mapped drive then it can't be used as a location of a database file.....

[dba_essex]

This is a local disk

[jmodic]

Yes, I concluded this from your later threads, realising that you have FailSave configuration. At first seeing partiton marked with leter as high as U: looked suspicious, but when I saw this is clustered configuration it made more sence.

As I see from one of your previous posts, you have two separate batch files - one for sutdown+file_copy and another one for startup. It is also very sympthomatic that you have this error reported on your index tablespace, which is backed up as the last tablespace in your backup script. Is it possible that you are trying to put the database cluster group back online before all the OS copy commands are finished in the backup batch? To be sure that this is not happening, why don't you put the whole process in one batch (or call both batch files from within the third, main batch) so that you are sure that everything is done in a serial mode?

[dba_essex]

Hi Jurij

Very good point. I also thought this might be the problem, so I have already changed the scheduled job to bring the resource back online 20 minutes after the database shutdown + copying of datafiles instead of allowing 10 minutes. I'll have to monitor the database this weekend.

Thanks for your help.

[dba_essex]

No O/S errors this weekend. I presume it was because all the files were copied over completely before bringing the resource back online.

Thanks for your help.