I’m about to setup my first TrueNAS Scale system for my homelab, once the ordered system and drives arrive. Up to now I have been using a Synology NAS with a 6-disk pool configured for SHR-2 (Synology Hybrid RAID with 2-disk fault tolerance), which is essentially RAID-6, so this will be my first foray into ZFS. So far I’ve done quite a bit of reading up on ZFS and TrueNAS (too much reading perhaps) and I have a few questions that I’ve seen discussed over the years, but I’m not sure what the current state (or advice) is on certain things.
FWIW my plan is to have a single raidz2 vdev with 4 20TB WD Red Pro HDDs. If I need more storage space, I’ll take advantage of the new raidz expansion feature and add another 20TB HDD.
So here are my questions:
- While I’m planning to have four 20TB WD Red Pro HDDs in my zpool, I will actually start out with three 20TB WD Red Pro HDDs and a single 14TB WD HDD, which should give me around 28TB storage space. In a month or two I want to swap out the 14TB HDD for a forth 20TB WD Red Pro HDD, so that I can get 40TB storage space.
- Am I correct in assuming that - once the fourth 20TB HDD is integrated into the raidz2 pool and resilvering has completed - the storage space of my zpool can be easily increased from 28TB to 40TB (or perhaps that even happens automatically)?
- I understand that TrueNAS uses ashift=12, which means the sector size is 2^12 = 4096 bytes. That seems to work well with the HDDs that I’m planning to use (WD Red Pro 20TB), since those have a physical sector size of 4096 bytes ( Internal Drive Advanced Format 4k Sector Size Support and Information ). I do notice though (in the website linked above) that these drives have a logical sector size that emulates 512 bytes, but can be formatted to a size of 4096 bytes (through a feature called “Advanced Format (AF)”).
- Is there any advantage in changing that logical sector size on my HDDs from 512 to 4096 bytes (assuming ashift =12 (or higher))? Are there any disadvantages in doing so?
- I also understand that - in particular for some SSDs that use higher physical page sizes - it might be advantageous to have the zpool created with a higher “ashift” value, e.g. 13 for 8k page size or 14 for 16k page size.
-
Can I change the default ashift value on the TrueNAS UI, when creating a new zpool or do I need to create the zpool on the command line (and then later import it into TrueNAS), if I want to use a non-default value for ashift?
-
If I understand this correctly, there isn’t much disadvantage to use ashift=13 or ashift=14 even on an HDD zpool (other than perhaps some wasted space for small files). Is that correct or are there good reasons to stay away from such shenanigans?
Thanks so much in advance for clarifying these things for me.