Unraid Users: Are you using single or dual parity disks? Do you have strong feelings either way?

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

Markess

Well-Known Member
May 19, 2018
1,152
768
113
Northern California
Major overhaul of my system in progress and thought I'd ask the question.

If I have a good backup, then single parity ought to suffice, but I can see how dual parity could make things easier/more convenient if I lost a disk(s).

I'm currently using dual parity, 4TB each, and want to start increasing the size of my data disks from the current 3 & 4TB size,. But, for the price of a pair of slightly larger disks for dual parity, I could get a single much larger disk to allow for much larger disks in the array.

But, I can't decide on which way to go. So, more experienced and longer term users, what do you prefer?
 

i386

Well-Known Member
Mar 18, 2016
4,241
1,546
113
34
Germany
Im not an Unraid user, I'm running hardware raid with windows servers at home :D

The more parity disks the better, especially with hdds.
If you had one drive dropping out during online expansion (or rebuilding an array) leaving your array in a degraded/not optimal state, you will know why ._.

Rebuilding/expanding puts a lot of stress on all involved devices, increasing the risk that devices drop out the array. You fight against these chances with parity disks.
 
  • Like
Reactions: nikalai

IamSpartacus

Well-Known Member
Mar 14, 2016
2,516
650
113
Major overhaul of my system in progress and thought I'd ask the question.

If I have a good backup, then single parity ought to suffice, but I can see how dual parity could make things easier/more convenient if I lost a disk(s).

I'm currently using dual parity, 4TB each, and want to start increasing the size of my data disks from the current 3 & 4TB size,. But, for the price of a pair of slightly larger disks for dual parity, I could get a single much larger disk to allow for much larger disks in the array.

But, I can't decide on which way to go. So, more experienced and longer term users, what do you prefer?
I ran single parity when I had 8total disks and now dual that Inhave 16 total disks. Thats really just a preference because I travel a lot and just want to ensure uptime if I'm away for long periods. But honestly, single parity is fine for many since the data on your disks is not striped. You obviously want proper backups for any irreplaceable data on your array regardless of how many parity disks.

Im not an Unraid user, I'm running hardware raid with windows servers at home :D

The more parity disks the better, especially with hdds.
If you had one drive dropping out during online expansion (or rebuilding an array) leaving your array in a degraded/not optimal state, you will know why ._.

Rebuilding/expanding puts a lot of stress on all involved devices, increasing the risk that devices drop out the array. You fight against these chances with parity disks.
Unraid is not like traditional raid as the data is not striped. So even if you lose data disks beyond how many parity drives you have, you only lose the data on those data disks you've lost. Furthermore, the parity rebuild process is much less intense. A 10TB rebuild for example only takes me about 17hrs.
 

Markess

Well-Known Member
May 19, 2018
1,152
768
113
Northern California
I ran single parity when I had 8total disks and now dual that Inhave 16 total disks. Thats really just a preference because I travel a lot and just want to ensure uptime if I'm away for long periods. But honestly, single parity is fine for many since the data on your disks is not striped. You obviously want proper backups for any irreplaceable data on your array regardless of how many parity disks.

Thanks! Currently, I've got 2x4TB providing Parity for three data disks (1x4TB and 2x3TB) that are only around 30% utilization, with files and backups of the family's PCs, which I suppose is either overkill, or paranoia, or both :cool:.

I'm finally going to get serious about digitizing the family's DVD/Blueray collection though, and since the Unraid server is the only machine that's on all day, I'm going to put it all there. That means more data disks to store less critical data and need to decide if I want to buy more 4TB data disks, or get larger parity disk(s) instead (and move the current parity disks over for data).

I'm leaning toward the latter, just need to decide on one or two. SInce the bulk of the data is going to be less critical and largely static, I was thinking I could easily get by with a single parity disk. Love Unraid though for giving me that option in the first place!
 

IamSpartacus

Well-Known Member
Mar 14, 2016
2,516
650
113
I'd definitely go with the larger disks if you can swing it because you'll be really setting yourself up much better for future upgrades so that you don't have to upgrade parity again for a long time.
 
  • Like
Reactions: Markess

Spartacus

Well-Known Member
May 27, 2019
788
328
63
Austin, TX
I'm running dual parity and have been since the beginning when I only had 4 drives, I prefer to have a secondary disk fallback while parity is rebuilding if that occurs.
All of my drives are the 8TB WD white shucks, but I upgraded the parity drives to SAS HGST drives (also 8TB) reduced the monthly parity check from about 20 hours to 16 hours (7200 rpm drive and faster r/w overall vs the wd white shucks).
Sitting at 10+2 drives currently and about to add an 11th.

The best buy wd whites are my go to as they regularly $140 and $120 or $130 a couple of times a year (I like the BB points I get and I'm elite plus because of all the drives I bought :D).

Personally I'd like to see a third parity disk implemented into unraid.
 
  • Like
Reactions: Markess

IamSpartacus

Well-Known Member
Mar 14, 2016
2,516
650
113
I'm running dual parity and have been since the beginning when I only had 4 drives, I prefer to have a secondary disk fallback while parity is rebuilding if that occurs.
All of my drives are the 8TB WD white shucks, but I upgraded the parity drives to SAS HGST drives (also 8TB) reduced the monthly parity check from about 20 hours to 16 hours (7200 rpm drive and faster r/w overall vs the wd white shucks).
Sitting at 10+2 drives currently and about to add an 11th.

The best buy wd whites are my go to as they regularly $140 and $120 or $130 a couple of times a year (I like the BB points I get and I'm elite plus because of all the drives I bought :D).

Personally I'd like to see a third parity disk implemented into unraid.
Do you not have backups? I dont forsee them implementing triple parity any time soon.
 

Markess

Well-Known Member
May 19, 2018
1,152
768
113
Northern California
I’m in IT though, I’m overly paranoid.
Yes. This, has been my normal operating state, since I first got into computers a while (a long while) back.

Plus, I add a thick layer of "I'm lazy" like icing on the top. I know that RAID isn't a backup, but recovering from a lost disk via an array seems less labor intensive than restoring from a backup.

On the other hand, I haven't lost a mechanical disk in....15 years? But, maybe that just means I'm long overdue :eek:
 

IamSpartacus

Well-Known Member
Mar 14, 2016
2,516
650
113
Oh I do, I have a full secondary redundant array, I’m in IT though, I’m overly paranoid. Just not quite paranoid enough to want to do raid 1.
Paranoid about what though? It seems you've already successfully solved your paranoia by having a second local copy of everything and a cloud copy of everything priceless (presumably). So what does having a third parity drive on 1 of your two identical arrays get you? All it's adding as greater potential uptime in the rare case that you would lose 3 drives before a successful data rebuild. And even if you did, you have a local backup you could easily just copy your data back over quite quickly.

Your best bet is to just have a warm/cold spare drive or two that are already pre-cleared so that if you do lose a drive you can immediately start a rebuild.
 

Spartacus

Well-Known Member
May 27, 2019
788
328
63
Austin, TX
Well part of it is I’m lazy but on top of that I’m cheap, I have a drive with 48 bad sectors marked and its been increasing every few weeks. Too cheap to replace til it fully fails (be it may I bought a drive and have it precleared and ready to immediately replace). I plan to get up to 20 drives eventually I would like three of those to be parity ideally as my preferred ratio is 6 to 1.
 

Markess

Well-Known Member
May 19, 2018
1,152
768
113
Northern California
Your best bet is to just have a warm/cold spare drive or two that are already pre-cleared so that if you do lose a drive you can immediately start a rebuild.
That might be the most effective solution for me. One of my data drives (the newest...of course) turns out to be SMR. At least its only one. I was planning to pull it as part of this update, but can clear it and keep it as a spare. It appears from the Unraid forum that the current state of spin-down support for spares is that you can keep them spun down with the Unassigned Devices plugin or a script? Doesn't look like they get automatically spun down like the array devices do?

Lazy and paranoid...a very dangerous combination.
Yeah. It definitely has had its drawbacks for me at least!
 

Tom5051

Active Member
Jan 18, 2017
359
79
28
46
hardware raid 6 with hot spare here - 20x 2TB drives. 96,000+ hours on the drives. Have just started to replace them with new 2TB drives, would have preferred less of and larger drives but I'm on a budget
.
 

IamSpartacus

Well-Known Member
Mar 14, 2016
2,516
650
113
hardware raid 6 with hot spare here - 20x 2TB drives. 96,000+ hours on the drives. Have just started to replace them with new 2TB drives, would have preferred less of and larger drives but I'm on a budget
.

Not sure how this comment is relevant to a thread about Unraid.
 

Spartacus

Well-Known Member
May 27, 2019
788
328
63
Austin, TX
Your best bet is to just have a warm/cold spare drive or two that are already pre-cleared so that if you do lose a drive you can immediately start a rebuild.
I have a warm spare setup in both the backup and main box in case of failure(2 on the backup box since thats where the high sector drive is). I just tend to wait until true failure notation rather than warning because of the redundancies I have in place. Calculated minimal risk :).

Lazy and paranoid...a very dangerous combination.
Also the general mantra of IT, "World class IT....on a budget" also, “Thats dangerous we need to fix that asap...tomorrow” :D