[ Disclaimer: I had finished my post and got an error regarding number of images, which is limited to one item for new users, though I’ve had extended rights in the old community forum. So sorry for the external link (which is also limited, therefore the global album link), but those pictures are relevant for the post. I’ll gladly update my post with embedded images as soon as I’m allowed by the mods. ]
Hi everyone,
after finalizing and setting up my first TrueNAS custom build (see discussion here), I’m now running TrueNAS SCALE and so far everything seems to be working as it should be - although the system is not yet in full-production use.
However, as this is my first TrueNAS system, I still have some questions and also wanted to ask if you could take a look at the current configuration, basically a “sanity check” so to say.
As described in the other thread, I have
- 1x OS drive
- 2x Optane SSDs
- 4x Western Digital HDDs
- 4x Toshiba HDDs
Those eight conventional drives span four 2-way mirrors (so similar to a RAID10 set). I have taken care that each vdev has one HDD from each of the two manufacturers. The two Optane drives are configured as a mirrored SLOG volume.
<screenshot "VDEV layout">
I have already set up a few SMB and NFS shares (though this does not represent the full final setup) and rsync’d the corresponding data to the datasets. Datasets holding large files (e.g. “Photos”) are configured with a record size of 1M and all others with the default of 128KiB. Sync is set to ‘always’ for reasonable datasets, all others are left with the ‘standard’ default.
<screenshot "Datasets">
E-Mail alerts are configured and tested, Prometheus/InfluxDB integration will follow in the coming days. Regarding recurring tasks, I have set up three scheduled tasks (besides an additional cronjob for the “multi_report.sh” script each Monday morning).
- S.M.A.R.T. (short) – daily at 0:00
- S.M.A.R.T. (long/extended) – bi-weekly on the 10th and 24th day at 02:00
- Scrub – bi-weekly on the 4th and 18th day at 02:00
<screenshot "Data Protection Tasks">
<screenshot "Cron Jobs">
Unfortunately, it doesn’t seem to be possible to select the disks by UUID for the S.M.A.R.T. tests, as I would have liked to stagger those by manufacturer (so, for example, starting with the four WDs and two days later the remaining four Toshiba drives). I know that you can select the sd* descriptors in the GUI for this, but those are way too random regarding reordering after reboot, so as long as there is no hidden magic behind this GUI selection (e.g. actually selecting the drives by their UUID), this is not an option.
Should it be of interest, I also have extracted some statistical data for the last week (the peaks represent the times when data has been copied over to the datasets):
CPU/Disk/Memory/ZFS stats
<screenshot "Stats/CPU">
<screenshot "Stats/Disks">
<screenshot "Stats/Memory">
<screenshot "Stats/ZFS-a">
<screenshot "Stats/ZFS-b">
So, having this as a primer, I have the following questions:
- Is there anything set up entirely wrong or have I forgotten something essential?
- Is the S.M.A.R.T. and scrub schedule reasonable as it is configured right now?
- Do you have any recommendation on a surefire way to sync all the data to a QNAP box on a per-dataset basis?
- How would you automate the TrueNAS configuration backup (as the GUI only provides a manual “one-shot download”)?
Many thanks for your help so far in the initial thread and of course for any help and recommendation coming in on here! 