FM Server crash during backup

[daniel z]

Hi all

Yesterday, during a routine scheduled backup, my filemaker server 5.5 decided to crash - see attached pics.

We have a backup schedule that writes backups every hour, and this is rotated between 3 different locations so there will always be 3 backups of the work done in the last 3 hours. This has been working fine until now - for years.

Looking through the event logs, I saw that the crash occurred while the server was trying to backup the largest fp5 file which is 270MB in size.

Any ideas about why this could have happened ? My sysadmin colleagues assure me that the hardware and network setup are not at fault and are all good quality, although I have one suspicion that the server which stores one of the backups is being used for file serving as well and this may put it under more strain ?

Any recommendations about how backups should be performed ?

Thank you in advance

Daniel

[Wim Decorte]

I'd go over the other logs in some detail to see what kind of activity was happening around that time.

There really are only two ways of taking backups:

- through the FMS scheduler

- or using the Windows scheduler with a batch file that uses the FMS command line

But both use the exact same mechanism.

Make absolutely sure that no other process or user is touching the live fm files (no virus scanner, n o indexing, no user trying to copy the files,...)

[daniel z]

Thanks, Wim.

Could a damaged / corrupt database file be responsible ?

Dan

[Wim Decorte]

It could... but if the thing was rock steady for such a long time then something must have happened recently. Depending on how you've set up the log files you may be able to go back far enough and find something. Power out? Some of the hardware reporting an imminent failure?

[NYPoke]

For anybody still interested, we had the same problem. It got really bad at one point. We have 3 FM 5.5 Servers.

Our primary Server crashed several times last May. We gave it a new, dedicated machine, which did not help at all.

Eventually, we came up with two sources to the problem:

1) Client Usage during the backup. If the users continue heavy client use during the backup, the problem gets worse. When we made everybody take a break, the system did not crash.

However, that did not fix the problem, just hid it for a while.

2) I'm not sure who recommended this (maybe New Millennium), but we cleaned up out files, especially old file references.

Once we cleaned out the unused scripts, fields, layouts & file references, the Server Crash Problem subsided. I think we have only had one crash in the past 6-8 months, and we don't make our people take breaks anymore.

[daniel z]

I am still interested ! Thanks for the update.

Dan