Solaris agent 2.0.0.0126-sparc : multiple jobs at same time ?

[ThierryF]

Hello,

Assuming you have 2 folders to backup on a Solaris node acting as file server for Oracle db servers ...
(in real life, I do have 65+ folders)

As folder 1 and 2 receive files at different moment from different hosts, both cannot be backuped at same
time to preserve data consistency.

So, I created a job for folder1, and another one for folder2.

I start job#1 ok. then,when #1 is completed, job#2 ok. Both are ok.

Starting #2 when #1 is running fails with "Error: Another backup job is running."

This is a very blocking problem for us : as folder #1 and #2 are used to store db backups, I have no backup window
without active db backup jobs, knowing more than 65 are used ... the idea was, as being setup currently with networker,
when backup job is done, to start veeam backup job from cli. This is just ... impossible with Veeam.

Any Idea ?

Th

[PTide]

Hi,

In other words, you need the agent to be able to run jobs concurrently? Installing more agents on different servers (that have access to the same folders) is not an option, right?

Thanks!

[ThierryF]

No, not possible.

In my case I should have 65 solaris servers or solaris zones to backup a single filesystem of 8TB...
Next, time it will take to dump folder will require no file modification to keep data coherency of
each db backups stored on filesystem.

I also tought taking zfs snapshot before launching veeam backup but as nearly all dbs backups being daily full,
backup chain rebuild will take age to complete, generating unbelievable IOs ...
(7* 8TB backups to regenerate a new 8TB one) ...

Having backup started from different nodes is so not an option.
This is a particular case of backups but you may have business needs
to initiate backup of different folders, at different times, from same agent.

A Database server, for sample, powering different dbs...
Oracle Archives being generated at a different rate according db usage,
you may want to dump each Oracle Archives dynamically (unscheduled job but
a job started manually) so you may face concurrent jobs from same server.

By extention, you may have VMs from different jobs, with different backuped drive/disk/exclusion settings,
to run concurrently. I haven't tested it because not faced the isue yet but I feel it will also fail.

Thanks !

Th