TrueNAS 12.1U1: Confused about zfs, raidz and how to integrate new discs.

[stakanov]

Ok, I am a newbie coming from the Linux world. I started to use TrueNas shorty ago "just to have a try". Now it does what I want (miniDLNA over ethernet to my parents TV to make them have movies without add trash on demand).
While when I did set the system up I used what I had as electronic trash at hand, I did not care about quality. So while the Phenom AMD and the AMD3 mainboard with 8GB are ok for the task, the discs appear now a bold gamble, when it comes to survival time.
I first did set up the system without knowing anything about FreeBSD and ZFS and therefore I just did let "zfs" do. I know what zfs does in principle, as I do understand (in principle) RAIDz, RAIDz2 and RAIDz3 and limitations. I checked either the tutorial about ZFS as well as the RAIDcalculator.

For what I understand I have currently a "stripe volume" as vdev, made of two very(!) old 290 GB SATA and one 4TB Western Digital 5400 consumer disk for the first pool and a single disc for the second. I read und understood that the speed will be limited to the slowest disk in this compound, whether stripe, mirrored data or RAIDzn.
I allow myself for clearness the outlay of the discs:

# camcontrol devlist
<WDC WD40EZRZ-75GXCB0 80.00A80> at scbus0 target 0 lun 0 (pass0,ada0)

<ST3320620AS 3.AAJ> at scbus1 target 0 lun 0 (pass1,ada1)

<ASUS DRW-1814BLT 1.14> at scbus3 target 0 lun 0 (pass2,cd0)

<WDC WD2500BEVT-08A23T1 02.01A02> at scbus4 target 0 lun 0 (pass3,ada2)

<CT240BX500SSD1 M6CR013> at scbus5 target 0 lun 0 (pass4,ada3)

<WDC WD4003FFBX-68MU3N0 83.00A83> at scbus6 target 0 lun 0 (pass5,ada4)

<Hitachi HTS541612J9SA00 SBDOC70P> at scbus7 target 0 lun 0 (pass6,ada5)


<Generic USB SD Reader 1.00> at scbus11 target 0 lun 0 (da0,pass7)
<Generic USB CF Reader 1.01> at scbus11 target 0 lun 1 (da1,pass8)
<Generic USB SM Reader 1.02> at scbus11 target 0 lun 2 (da2,pass9)


While an imported command (from Linux) shows:
# lsblk
DEVICE MAJ:MIN SIZE TYPE LABEL MOUNT
ada0 0:115 3.6T GPT - -
<FREE> -:- 44K - - -
ada0p1 0:129 8.0G freebsd-swap - -
ada0p2 0:130 3.6T freebsd-zfs gptid/e3ab84fa-3d79-11eb-9ad8-00241d7dd2a4 <ZFS>
ada1 0:116 298G GPT - -
<FREE> -:- 44K - - -
ada1p1 0:131 8.0G freebsd-swap - -
ada1p2 0:132 290G freebsd-zfs gptid/a3b03f72-1629-11eb-be34-00241d7dd2a4 <ZFS>
ada2 0:125 233G GPT - -
<FREE> -:- 44K - - -
ada2p1 0:133 8.0G freebsd-swap - -
ada2p2 0:134 225G freebsd-zfs gptid/6753e4c2-162b-11eb-be34-00241d7dd2a4 <ZFS>
ada3 0:126 224G GPT - -
ada3p1 0:135 512K freebsd-boot gptid/6b782e7d-93ac-11ea-bc74-00241d7dd2a4 -
ada3p3 0:137 16G freebsd-swap gptid/6b7c0447-93ac-11ea-bc74-00241d7dd2a4 -
ada3p2 0:136 208G freebsd-zfs gptid/6b7f49ac-93ac-11ea-bc74-00241d7dd2a4 <ZFS>
<FREE> -:- 8.0M - - -
ada4 0:127 3.6T - - -
ada5 0:128 112G GPT - -
<FREE> -:- 44K - - -
ada5p1 0:138 8.0G freebsd-swap - -
ada5p2 0:139 104G freebsd-zfs gptid/8f158273-162b-11eb-be34-00241d7dd2a4 <ZFS>
da0 0:144 - - - -
da1 0:145 - - - -
da2 0:146 - - - -

(I have also the "gpart show" but is largely equivalent and IMO less clear).
Where root is red, the new, unformatted disc (that will be joined by a second new unformatted of 4TB in few days) is green and the one 4TB to "stay" is blue.

Now I hope that was confusing enough.
I have two pools. One striped called Atlantico with ada0, ada1 and ada2. This is the one I wish to "bloat". The other (which is currently a single disc and is not meant to change in the moment is Andaman and is done of ada5.

Now once the second 4TB disc arrives:

is it advisable to create a "mirrored" filesystem or RAIDz1 with three 4TB disc?

(The 2 smaller discs will be removed and attributed eventually afterwards to use with Andaman as "mirrored" with two discs).

If, creating a RAIDz1 or a mirrored set:
Do I have to rebuild the pool (saving the data outside in the meanwhile) or can I "offline" one after the other the small discs (replace disc) and then join the new 4TB?
I suppose I will have to destroy the pool to have a RAID, but suppose does not mean "have understood".
If I create this, should I "mirror" as described in the tutoria (mirror with three disc) or do RAIDz1 - if the idea is redundancy - ?
If afterwards I am rich and fortunate and can buy two more 4TB discs, can I change from RAIDz1 to RAIDz2 once I have the disc or is it again: save outside and then destroy and rebuild?

Sorry for the very basic questions, but the more I read about zfs the more I am getting confused.
Thanks in advance.

 

[stakanov]

In the meanwhile (now answers) I did rebuild all, with a raidz2. So consider this closed.