System cannot boot after backup fails

[Bryan Fahy]

Hi there
I do not have exact error message from customer but it is a standard console message realating to boot halt after a backup on Tru64 has failed.

Backup fails
System defaults to try to run the job again
(display on console is something like - Run the job again - YES or NO)It is necessary to manually choose NO and after this action, the system boots OK
Before manual intervention - The halt causes the networked system to be isolated from network


Question:
Do you know if there is a file that can be edited to automatically set NO as default for running the same backup job again - this will prevent system HALT??

[Michael Schulte zur Sur]

Hi,

what kind of backup are you using?

greetings,

Michael

[Kasper Hedensted]

Hi Bryan,

I think we worked together a few years ago.
Is this the Euro**** customer ?

Cheers,
Kasper

[Bryan Fahy]

Hi Kasper,

Good to hear from you again! What a co-incidence ! Yes - still working as Team Lead for EURO****.
I am moving onto the transition team shortly - after 3 years its about time for a change!

How are you?

Bryan

[Kasper Hedensted]

Hi Bryan,

I am fine thanks, hope that you are doing well.

If I remember correct they have a special backup procedure:

A cron-job touches a file and then reboots the system;
Then during boot a backup-script located in /etc/inittab looks for this file and if it exist, it runs a vdump-backup before it runs the rest of the startup scripts.

I can't remember if they run dump or vdump, but you can see that in the backupscript mentioned in /etc/inittab.

It would be interesting to know the exactly error-message, try and see if the backup-script logs anything.

The problem might be that the vdump command prompts for something, and the system does not boot until the backupscript exits.
Maybe it is asking for another tape or something similar.

I don't think that you can set it to automatic choose "NO". It prompt because an action is needed.
If it asks for another tape try and cleanup the filesystems, maybe some large files could be removed. Or maybe they could use larger tapes (if the streamer can handle it).
Another approch could be some sort of timer which makes sure that the backup-script cannot run forever (and logs that the backup has failed).

Cheers,
Kasper

[Bryan Fahy]

Hi Kasper,

Thanks for the info.

The backups are created with dump utility.

WE were also thinking of running "yes | dump -flags"

WE will do some work today on recreating this issue locally and testing the various suggestions. I will let you know the outcome

Talk soon
Bryan