FreeNAS Improve Timeouts When Activity?

[Devin Acosta]

I have FreeNAS installed a dedicated HP DL360g8 Server that has 3 x 10TB drives in a RAID-Z1 configuration, and I have an SSD as cache. I do have De-duplication turned on as I'm copying nightly system images to the appliance. The Server has 128GB of memory and the Storage Pool has 15TB of usable storage. I know probably having de-duplication on is probably mostly my problem but I was trying to save as much storage as possible. I notice that when i'm copying the nightly backups to the appliance or i'm doing my nightly purge job of deleting old backups, that during this time any box that has the appliance mounted times out, and the kernel complains that the NFS server went away for a period of time.

Is there anything that I can do to at least make it not timeout or respond slowly to NFS requests? I presume either my IOPs on the few disks that I have just isn't cutting it and I may need to add a lot more drives. Just curious of other peoples thoughts?

 

[Chris Moore]

that has 3 x 10TB drives in a RAID-Z1 configuration

First, you need to do some reading:

Updated Forum Rules 4/11/17
https://forums.freenas.org/index.php?threads/updated-forum-rules-4-11-17.45124/

Slideshow explaining VDev, zpool, ZIL and L2ARC
https://forums.freenas.org/index.ph...ning-vdev-zpool-zil-and-l2arc-for-noobs.7775/

Terminology and Abbreviations Primer
https://forums.freenas.org/index.php?threads/terminology-and-abbreviations-primer.28174/

and I have an SSD as cache

Why? What kind of cache?

I notice that when i'm copying the nightly backups to the appliance or i'm doing my nightly purge job of deleting old backups, that during this time any box that has the appliance mounted times out, and the kernel complains that the NFS server went away for a period of time.

Hard to say without any idea how your hardware is configured. There are far too many options for how to configure a HP DL360g8 server, so just saying that it has 128GB of memory and the Storage Pool has 15TB of usable storage doesn't tell enough. Just giving the model does not tell what you have in it. For instance, is it on a 10Gb network? Are the drives connected to a SAS controller.

I presume either my IOPs on the few disks that I have just isn't cutting it and I may need to add a lot more drives. Just curious of other peoples thoughts?

As nice as it would be to know more, it is most likely that you don't have enough drives. You shouldn't be running RAIDz1 with 10TB drives to begin with, but aside from that, if it is speed you are looking for, you need a lot of drives, not big ones. To give you an example, one of the backup systems I have uses 80 drives that are 4TB each and I have them broken into 6 drive vdevs because vdevs give you IOPS. The more vdevs you have, the more IOPS. In a single vdev, as you have it, you are only getting the speed of a single drive. If you are trying to hit that drive with traffic from multiple systems at the same time, the drive just can't keep up.
The Sun / Oracle SAN in our main data-center has 224 spinning disks and all of that is about providing the IOPS because the individual drives are only around 1 or 2TB if I recall correctly. I am not saying that you need 200 drives, but it is safe to say 3 is not enough, since you are having a timeout.