Is this a bug in 3.4?

[coelhu]

Hello

I have noticed some strange behavior in version 3.4

amdump does not check if the job is running and start's it all over again ignoring the running job, older versions of amdump used to give a error that the job is already running and this is the expected behavior.

Is this fixed in versions above 3.4?

I could not install versions above 3.4 because strange ssh errors, but for this erros I will open another thread.


Thanks

[martineau]

It is a new feature, 3.4 allow multiple backup run in parallel.
Why is it important to not allow multiple run in parallel?

[coelhu]

Multiple backups for the same job? If the job currently running is a full backup that have taken too long and the cronjob executes another dump how amanda knows what to make backup if the full is not yet finished?

There are some jobs that take one day or two to finish the full backup and while they are executing is very system consuming have the same job running again...

Is there anyway to disable multiple backup run of the same job?

Or my backup strategy should be different?

[martineau]

You can write an amdump wrapper like:
{code}
#!/bin/bash

if [!-e /etc/amanda/CONF/running]; then
touch /etc/amanda/CONF/running
amdump CONF ...
rm /etc/amanda/CONF/running
else
echo "already running"
fi
{code}

[coelhu]

OK, this solution is nice. Thanks.