Full System Restore Failed ???

[antzzz]

My colleague accidentally initialized the disk using oninit -i on our Standby server (IDS 7.3 on SCO Unix 5.05). The chunks (ie raw devices) are still there so we just recreated the dbspaces and added these chunks exactly as the Live one. We tried to do a full system restore on this server using a backup from the LIve one but it keeps failing everytime. The error is as follows (during an ontape -r) after answering the question if I want to backup the logs:

Physical restore failed - function read archive backup failed code -2 errno 0

Now we've checked everyhting: permissions (660), using user Informix, same onconfig on both machines, etc.

Can anyone help?

P.S. Do you need to update the Reserved Pages on the Standby server after you add dbspaces (ie similar to adding new logs) and do an ontape pointing to /dev/null? We haven't done this yet. (ie the server is empty, right? and the backup is from a different machine)

 

[ifxtechie]

Hi,
Pls make sure that the configuration shown by ontape and the current configuration [onstat -d] should match. See that all the DB-Spaces are same though some may be cooked files.
You need to compare each and every db-spaces and its type.
If any of the db-spaces is absent, then add the dbspace and try restoring.
Here I would suggest some things which you MIGHT HAVE DONE.
1. I feel from your notes that the tape drive is working.
If not, make sure that the tape drive is working. This is done by tarring some files to the tape and extracting those files onto the drive. If not working, try cleaning it. If still not working, then you need to change the drive. [it is the responsibility of SyStem Administrator]
2. If the tape drive is working then make sure about the configurations of DB-spaces. [report of onstat -d and the info shown as soon as you hit ontape -r will help in this step.]
3. Another important step is make sure about the block size. This should be equal or grater than that of backup.
Pls let me know if this doesnot solve the problem.
Regards
Partha

 

[antzzz]

Thanks Partha for your help. Our tape device is pointing to a file (eg /backup/current.bck) same as the live one. We have actually successfully restored using this style months before. But this is the first time we are restoring to an "empty" database server. The onconfig on both machines are the same: block size, shared memory params, etc. We even recreated the dbspaces with the same chunk IDs as the Live one's with the same size KBs. The only thing that we didn't do was to do an ontape pointing to /dev/null to update the Reserved Pages after we added the DBspaces using onmonitor (do you need to do this???). We immediately tried to do an ontape -r after adding and recreating the dbspaces. We contacted (IBM) Informix and they said it might be the permissions of the chunks but these are the same as the Live one too (660). We have tried using both user Informix and root but the same error keeps popping up.

 

[schillache]

Hi,

1) After this error happens,u just see /var/adm/messages files whether is any problem in Tape media.
2)After this error happens.U just do onmoke -ky twice(for safe)
then oninit & rerun onmode -ky & run ontape -r.

try this & let me know the status.

rgds
schillache