live backup problem

[VolkerBarth]

hi all,

i have a problem with live backup. i run live-backup as a windows task (at system startup). The task runs for about 8 hours and then quits without any error. The return code is (5) [interrupted by user]. i also run full-backups every hour. after the 8th full-backup is done, the live-backup quits. if i run full-backup every half an hour, the live backup quits after the 16th backup. The live-backup-task also quits when running on cmd (without sched. Task). The system doesnt run out of disk space. Can anybody help me?

[VolkerBarth]

What version do you use? Is the live backup process running on the same machine as the database engine? When using "DBBACKUP -l ... -o MyLiveBackupLog.txt" what does MyLiveBackupLog.txt show afterwards?

You could also add the "LOG=" connection parameter to the dbbackup -c option to debug the connection between the engine and the live backup client.

[Former Member]

i am using version 12; live backup is NOT running on the same machine. The log: Transaction log truncated by backup -- restarting ... (6385 of estimated 6385 pages, 100% complete) Transaction log truncated by backup -- restarting ... Database backup interrupted by user

I try to use the LOG= parameter thank you so far.

[Breck_Carter]

Is there anything in the event log?

FWIW one method to deal with people accidentally stopping the task was to start live backup inside a never-ending loop in a batch file. The attempt to start the backup would fail repeatedly until it was needed :)... I misremember the exact details since this was from the days of V9.

[justin_willey]

We have a windows service that checks that dbbackup is running and restarts it if not.

[Former Member]

because i am still in testing i forgot to delete the livelog file. so i have to wait another 8 hours. if you wonder why i dont use dbbackup with -y ?!? because i have another server which does a live backup of a third server without -y running. This live Backup runs for 3 month..... everything is exactly the same. Two things are not: The -y option and the user which is running the task. On the working server i run it with the domain-admin, and on the not-running-server i run it with a "normal" user which is also member of the domainadmin group...

[Former Member]

can you tell me how to create this service?

[Former Member]

event Log:

05:44:38 IP=10.8.1.71;HOST=DE08SNUMBACKUP;OSUSER=Administrator;OS='Windows 2008R2 Build 7601 Service Pack 1';EXE='C:\\Program Files (x86)\\SQL Anywhere 12\\BIN32\\dbbackup.exe';PID=0x318;THREAD=0x1494;VERSION=12.0.1.4231;API=DBLIB;TIMEZONEADJUSTMENT=60
05:44:38 Connected to the server, attempting to connect to a running database...
05:44:38 [  541] Connected to database successfully
13:35:54 [  541] Client disconnected
13:35:54 [  541] Disconnected from server

[justin_willey]

We wrote it in Delphi, but you could use any programming language that will let you create a Windows service.

[Breck_Carter]

That's the SQL Anywhere server console log. I was asking about the Windows event log... are there any pertinent entries around the time dbbackup stopped?

[MarkCulp]

In addition to Breck's request for info from the server console log, I would ask that you look at the computer's event log on the computer that is running dbbackup.

I looked at the source for dbbackup and traced the source of the "Interrupted by user" message. It would appear that something on the computer running dbbackup sent a SIGINT (or SIGBREAK) to dbbackup. So I would look for anything that might be wanting to stop the dbbackup program. E.g. Do you have some monitor process that might be killing programs that run too long? (where 'too long' here would be about 8 hours).

[Former Member]

in the eventlog i only get messages like: Task Scheduler successfully completed task "\\Live_Backup_prod" , instance "{33ef979f-dea9-4278-84f3-4ddb56e0b415}" , action "C:\\Windows\\SYSTEM32\\cmd.exe" with return code 5.

and

Task Scheduler successfully finished "{33ef979f-dea9-4278-84f3-4ddb56e0b415}" instance of the "\\Live_Backup_prod" task for user "Domain\\Administrator".

looks like the task is configured to run for 8 hours but it isnt.

it is a absolutely clean installed server Nothing but sqlanywhere is running on it. Like my other server which runs the dbbackup for weeks.... same operating system, same gpos, but there a 2 differences: - machine A (which runs backups for weeks) is a vm machine B is a physical server

• machine B has 4 physical network interfaces (teamed) if a network interruption is the problem, i assume the return code should not be 5

[VolkerBarth]

I guess it's not the server console log but DBBACKUP's output log with the "LOG=" connection parameter, right?