Hi!

 

unRAID newbie here, that has just gotten around setting up my first unRAID NAS.

 

The current setup:

- HPE ProLiant Microserver Gen8, Intel Celeron G1610T, 4GB of RAM.

- 2 x brand new 4TB Seagate IronWolf HDD (ST4000VN008), never been used or connected to anything else before this setup. The plan is to add another 2 identical disks later, to fill up all 4 HDD bays.

 

The problem/question I'm having is about the initial parity sync, since I think that it is going at a pretty slow pace. As you can see in the screenshot, I have setup one of the two disks as a parity drive, and is performing the initial parity sync. Is 30-35MB/sec normal in this case, or is there something that I should be doing differently? It have been staying around that speed since the start yesterday.

 

 

The Microserver Gen7 has a write cache bios setting (disable by default) that gives similar performance, don't know if the Gen8 has a similar setting.

The Microserver Gen7 has a write cache bios setting (disable by default) that gives similar performance, don't know if the Gen8 has a similar setting.

 

Hmm, I'm not sure about that, I'll have to check! However, I think that I'll wait until the parity sync is complete, I would hate to have waited this long to abort the sync! 

 

Related question: When I later add 2 new drives, will I have to do this process again (parity sync)? And, as it is for now, could I expect similar speeds then?

No need to do a new sync, but the disks will be cleared before they're added to the array to maintain parity, process takes a few hours (at normal speed) but array remains online.

 

Alternatively can also use the preclear plugin, in this case more for disk testing.

No need to do a new sync, but the disks will be cleared before they're added to the array to maintain parity, process takes a few hours (at normal speed) but array remains online.

 

Alternatively can also use the preclear plugin, in this case more for disk testing.

 

Okay, thanks for the info!

 

Regarding the initial parity sync speed, what is considered to be a "normal" speed, for identical disks in terms of performance?

With modern disks it should start @ ~200MB/s for 7200rpm disks and ~150MB/s for 5400rpm, speed will progressivly decrease as it goes to the slower inner cylinders and end at ~100 and ~~80MB/s respectively, this when there are no controller or other bottlenecks, like in that microserver with only 4 disks.

With modern disks it should start @ ~200MB/s for 7200rpm disks and ~150MB/s for 5400rpm, speed will progressivly decrease as it goes to the slower inner cylinders and end at ~100 and ~~80MB/s respectively, this when there are no controller or other bottlenecks, like in that microserver with only 4 disks.

 

Hmm, okay, so then I'm correct in assuming that the speeds should be greater than what I'm getting? I'll let this parity sync complete first, then I'll have to check for that write cache setting in BIOS, and see if that could improve something. Btw, is parity check something that I could test the speeds with after the sync have completed? Also, is there any harm in starting a parity check, and then stopping the check?

Parity check speed usually = parity sync speed (though in this case if it's the write cache setting, parity check will be much faster because there's no writing).

Parity check speed usually = parity sync speed (though in this case if it's the write cache setting, parity check will be much faster because there's no writing).

 

Aah, I see! So, if I have a write cache setting in the BIOS, that should be disabled, right? That would also improve performance when writing to the array later on?

 

You'll have to excuse me for all these question, I'm just trying to wrap my head around all this. 

If you have the write cache setting would should enable it to get better performance in any operation that includes writes, including parity sync and normal writes to the array.

 

That setting, at least on Gen7, exists and is disable by default for users without a UPS, it should be enable if there is a UPS, it's not recommend without one, but it's also not recommend to use unRAID (or any other NAS) without one.

If you have the write cache setting would should enable it to get better performance in any operation that includes writes, including parity sync and normal writes to the array.

 

That setting, at least on Gen7, exists and is disable by default for users without a UPS, it should be enable if there is a UPS, it's not recommend without one, but it's also not recommend to use unRAID (or any other NAS) without one.

 

Okay, thanks for the info, I'll have to try it when the sync completes!