NFS/NFS4 as Datastore in PBS v2.4-1 ?

[LooneyTunes]

Hi,

I have a NAS setup for NFS4, and configured it to allow PBS to access the shares. But where in PBS do I configure a NFS datastore? I says to use absolute path in the datastore config dialog, but nothing about protocol? Can someone please tell me what I need to do on the PBS side? I have tried to read all I have found, even asked chatGPT, which did not exactly nail it this time...

Thanks

 

[Dunuin]

I have a NAS setup for NFS4, and configured it to allow PBS to access the shares. But where in PBS do I configure a NFS datastore?

PBS is meant to be used with local disks only, so that is not supported by the webUI or CLI tools. You would need to mount that NFS share yourself using the CLI (for example by editing the fstab). You can then use the webUI to create a new datastore and point the path to your NFS shares mountpoint. Also make sure that the backup user (I think it was UID 34) got privileges to access/write to that NFS share.

 

[LooneyTunes]

PBS is meant to be used with local disks only, so that is not supported by the webUI or CLI tools. You would need to mount that NFS share yourself using the CLI (for example by editing the fstab). You can then use the webUI to create a new datastore and point the path to your NFS shares mountpoint. Also make sure that the backup user (I think it was UID 34) got privileges to access/write to that NFS share.

Ah... That explains a lot, thank you! I will try this

 

[Dunuin]

Your NAS will probably also be too slow in case you plan to have some TBs of backups (at least in case that NAS is using HDDs and not SSDs). HDDs can't handle all die IOPS needed for GC tasks and the network and NFS overhead will add latency reducing the IOPS performance even more. But might be fine if you just got a few hundred GBs of backups. Otherwise the GC task might run for days making your NAS basically unusable while it is running, as all the disks will be at 100% utilization trying to read/write millions over millions of small IO caused by reading/writing metadata.

 

[LooneyTunes]

Your NAS will probably also be too slow in case you plan to have some TBs of backups (at least in case that NAS is using HDDs and not SSDs). HDDs can't handle all die IOPS needed for GC tasks and the network and NFS overhead will add latency reducing the IOPS performance even more. But might be fine if you just got a few hundred GBs of backups. Otherwise the GC task might run for days making your NAS basically unusable while it is running, as all the disks will be at 100% utilization trying to read/write millions over millions of small IO caused by reading/writing metadata.

Yes, I am unsure of the exact amount, but should be fairly low, so at least worth a try Thanks for the heads-up!