Priority of pbs tasks

[leksand]

I discovered during the backup that if several nodes are selected in the backup schedule, then several backup tasks are created that start running simultaneously.

The storage goes up to 10 Gbps, but currently, a server with a virtual machine with pbs has a 1 Gbps channel. I plan to eliminate this bottleneck, but:

is it possible to make a switch that will allow you to select the mode in which tasks will be run in turn - first a task from the vm of the first node, after completing this task - a task from the vm of the second node, and so on?

 

[aaron]

is it possible to make a switch that will allow you to select the mode in which tasks will be run in turn - first a task from the vm of the first node, after completing this task - a task from the vm of the second node, and so on?

That is currently not possible. The backup jobs run independent on each node. What you could do is configure multiple backups jobs and limit them to each node, and then let them start at different times.

 

[Neobin]

For reference:
https://bugzilla.proxmox.com/show_bug.cgi?id=3086

 

[leksand]

That is currently not possible. The backup jobs run independent on each node. What you could do is configure multiple backups jobs and limit them to each node, and then let them start at different times.

Thanks !

A question about bandwidth limitation: is it possible only through editing configuration files or in the graphical interface too? Ideally, I would like to be able to set limits for each task

 

[aaron]

A question about bandwidth limitation: is it possible only through editing configuration files or in the graphical interface too? Ideally, I would like to be able to set limits for each task

AFAIK there is no option to set bandwidth limits per backup job at all.

Looks like it is possible, but not exposed in the GUI. So either set it afterwards via the API, for example:

pvesh set cluster/backup/{job id} --bwlimit {limit in kiB/s}

Or in the /etc/pve/jobs.cfg add a line

    bwlimit {limit in kiB/s}

in the correct job section.

Under Datacenter -> Options -> Bandwidth Limits -> restore/backup you can set a global limit.

 

[leksand]

AFAIK there is no option to set bandwidth limits per backup job at all.

Looks like it is possible, but not exposed in the GUI. So either set it afterwards via the API, for example:

pvesh set cluster/backup/{job id} --bwlimit {limit in kiB/s}

Or in the /etc/pve/jobs.cfg add a line

    bwlimit {limit in kiB/s}

in the correct job section.

Under Datacenter -> Options -> Bandwidth Limits -> restore/backup you can set a global limit.

Thank you, so far I have set the limit on all points to 400 Mib.

How to correctly calculate the optimal limit depending on the bandwidth of the channel (now 1GB)? Copying goes to the pbs located on one of the pve nodes, storage on the storage connected as a block device - that is, it receives data and sends it over one channel. In the future, I will allocate separate ports and a subnet for backup or transfer to a 10 GB network