Getting started with my first truenas but I cant decide how to configure my apps and boot pools.
i have 3 ssds, all 256gb, two sata and one nvme
at the moment I use an sata ssd as boot drive, the nvme ssd for apps and trying to decide if I should the last ssd for redundancy on the boot pool och apps pool.
From what I am reading neither is really vital, at least for me since uptime is not critical. I plan on having backups for OS config on my google drive or elsewhere and configs for apps on the storage pool.
my thoughts:
-I could sacrifice an slot for a future storage drive, get another ssd and have redundancy on both pools. (8x3.5’’ slots in chassi, only use 3 today, plan to fill up but will take years, sata ports is limiting factor)
-worried about losing nvme performance if using an sata ssd mirrored with and nvme ssd
-leaning towards redundancy in the boot pool since a failed boot pool feels harder to troubleshoot than an failed apps pool
-if i skip redundancy on both pools I could use an ssd for torrenting and seeing linux isos, but i havent really looked into that yet, not sure if needed or not
For a pool, if you have redundancy with mirror and one disk fails - the service is still up and running, and you just need to work on fixing the redundancy issue. Now in case of boot pool that means, in case of a failure you have no NAS and Apps.
If you are managing with config backup and restoring it - the service will have a downtime, until the OS is reinstalled and config is restored.
So it boils down to the final question : Do you mind/care/worry if your NAS and apps were unavailable for the period of time during such fixing.(in other words downtime). The answer to this question will differ from case to case and person to person - When you have an answer to it, that will give clarity on which route to take.
A failing boot drive generally shows up as crashing middleware or no NAS at all. Very easy.
Bring in a new drive, install anew, load the configuration file. Done.
Corruption in the app pool may get more complicated if you have to redo configuration manually.
In a home NAS (= not “business critical”, downtime is acceptable), none is really required to be redundant, but I’d go for redundant app pool over boot.
And so what?
Move the boot to NVMe, or bring in a second NVMe drive. 256 GB drives are cheap.
For those who believe, a home NAS does not really need redundancy , think about these scenarios also :
what if you are just away from home and trying to access it remotely
what if you are totally engaged with some activity and your family members need to access data
what if you had configured backups to the nas from other devices, and you do not have time to fix the nas boot issue immediately.
what if you are unfortunately in a not so healthy situation to fix it immediately and your family/home members need to access some data. (try explaining the member how to fix these at such situations…it will not be so much fun)
So, even though many of us may take it for granted thinking “its a home NAS, not business critical - so it can wait…”. Try to think like this, does down time matter or not…
I would rephrase it as, if your day to day access to the resources on the NAS is not a problem, then redundancy can wait.
In most of the case I have seen, things know when exactly you are away or unavailable… and it waits for that exact beautiful moment to fail. So now a days, in my opinion : if its not a lab or test system, redundancy is a must.
In my opinion, when there are some limitation, is just a matter to choice the less pain way.
That’s why for me i have choice redundant app pool…
In my experience (i’m a TN user from the end of 2023) i have got 2 issue from disks designed for apps (1 particoular disks that sometimes fails on reboot, causing the need of a small resilver… Gotta let that happen 2 times before replace It ) and 1 total fail from boot disk (died without any previous simpthone during night)… Still never had data loss, so no need to use backups.
I also consider that (having dayli snapshots of the app pool) the possibility to lose some hour of data on apps Is hundred way worst than restore a some-day old config file (provided automatically and free from multi report every week).
I also have a spare cheap NVME disk ready to be used as boot pool in case of needs, instead of let It use and degrade for the little advantage
Not sure, but even if you mirror your boot pool, for the Bios/UEFI it still is two disks, and depending on the type of failure might still try to boot from the wrong disk. So you need to be on site, so to speak, to fix the problem.
The first thing come to my mind Is that the phisical access need to change disk Is not always granted, despite change boot order can be possible from remote location. That’s for sure a case where redundant boot pool Is pretty usefull
In such a scenario where on-site maintenance is not possible, you need three boot devices in ZFS mirror + hardware RAID 1 (you read that well: Hardware RAID with ZFS!)
Which highlights that a mirrorred boot pool, in itself, is nearly useless if the BIOS does not handle failure properly.
No room for more nvme drives on my mitx board. But yeah, guessing the performance difference in real world scenarios is probably not noticeable anyway, and losing data is worse than losing performance anyway
thanks