Pluvichor problems, split from "Dealing with unclean shutdowns"

[Pluvichor]

Hello, I'm hoping someone can help me. I'm very new to Unraid, but about two weeks back I performed what I now understand to be an unclean shutdown on my server (it was for a super dumb reason - I couldn't figure out my server's ip to connect to it from my pc. i have since learned how to do that and now have it set to a static ip anyways) and since then I've been experiencing issues getting my server to have a proper parity check and recognize certain drives at all. I have two 20tb drives for parity so parity checks take forever anyways, but now trying to build parity seems to hang late in the process and freeze the system (I just experienced that and had to ungracefully shutdown the system again because the gui wasn't responding).

 

It was such a nightmare, I decided to reformat everything and try to start from scratch (this was a new server so I wasn't super attached to the data on it anyways). I tried New Config under Tools a few times to no new results, so I formatted all the drives and installed Unraid anew on my flash drive. That brings to me to now.

As an aside, one of my drives (Disk 2) was giving me issues being unmountable and not being able to be formatted, but I somehow just fixed that issue (I hope) by letting the array run without parity for a few minutes. 

I'm running another parity-check right now (I just realized, it starts as a "parity check" when I first start the array with parity drives, but when I check on it later after work it'll have changed to a "read check" or something... is that significant?) and it's giving me an "Unraid array error" warning - 3 disks with read errors... is that typical?

Any advice on how to proceed (assuming this latest parity check does what the last few have done and goes nowhere) would be greatly appreciated.

EDIT: Two things I noticed while the parity-sync is ongoing, not sure how relevant they might be:

Firstly, so far it's at  542925099 errors between the 3 drives. That seems like... a lot. 

Secondly, the shares disappeared. I can't see them under Dashboard or Shares, and my computer can't access the empty "media" share I'd created soon after I'd started the parity-sync.

I've uploaded a second diagnostic zip done now while it's parity-syncing in case that helps.

starryskynet-diagnostics-20220713-1839.zip

starryskynet-diagnostics-20220713-2237.zip

Edited July 14, 2022 by Pluvichor
New info

[JorgeB]

@Pluvichorsplit your post here since this isn't really about unclean shutdowns.

 

There are errors on disks 1 and 2 immediately after array start, probably not coincidental that both disks are of the same model, first thing to try is to update the LSI firmware to latest, the one you're using has known issues, if that doesn't help try connecting those two disks to a different controller, you can swap with another two.

[Pluvichor]

I see, thank you! I'll try updating the LSI firmware right now.

[Pluvichor]

Okay, I'm trucking along trying to follow along to the instructions on that page, but I'm getting this permission denied error on the last two steps:

 

root@StarrySkyNet:~# cd /lsi
root@StarrySkyNet:/lsi# chmod +s sas2flash
root@StarrySkyNet:/lsi# sas2flash -listall
bash: ./sas2flash: Permission denied
root@StarrySkyNet:/lsi# sas2flash -o -f 9207-8.bin
bash: ./sas2flash: Permission denied
root@StarrySkyNet:/lsi# 

 

Also, where in the diagnostic/gui can I see what version my LSI card is? I bought it off ebay and thought it came pre-flashed to the newest version.

Edited July 14, 2022 by Pluvichor

[JorgeB]

Jul 13 18:23:12 StarrySkyNet kernel: mpt2sas_cm0: LSISAS2308: FWVersion(20.00.06.00), ChipRevision(0x05), BiosVersion(07.27.01.01)

 

Not sure about the permissions issue, make sure you're following the instructions correctly.

[Pluvichor]

Okay, so I haven't had a chance to work on updating the LSI bios, but I tried something last night that now has me thinking the issue isn't the LSI card.

 

So I have 16 drives, half of them connected by sata to my motherboard, and the last 8 connected to (what I thought was an 9207-8i) an LSI card. Crucially, the 4~ drives giving me issues (2 parity and disks 1 and 2) were connected by sata to sas adapter cable to the LSI card. So I reconnected them from the LSI card to the motherboard sata ports. But it's still those four drives giving me issues and not the ones current connected to the LSI card now.

 

When I first put the unraid build together a few weeks past (and it was working without issue), these four drives were originally connected to the motherboard before I swapped them to a different area of the case for better Temps, and in doing so connected them to the LSI card. This happened prior to me having these issues I'm currently having so I didn't connect it before now, but now I'm looking at my setup and wondering if it's a power issue? I have something like 12 drives daisy chained together, and these drives are the last ones at the very end of the chain. They're also the ones running the most power I imagine because Parity are at the very end and Disk 1 held most (all?) of my data because of high-water.

 

Last night the server seemed especially... slow? Mounting and unmounted the array took a long time and when a ready-check started, it said it would take more than a year to finish. When I stopped the array, the parity disks and disks 1, 2 and 4 would randomly disappear from the list (again, coincidentally the last drives in the power daisy chain except for 3) as if unraid couldn't find them. I'm not sure how this is related to my issues if at all.

Edited July 14, 2022 by Pluvichor
Grammer

[JorgeB]

13 minutes ago, Pluvichor said:

I have something like 12 drives daisy chained together, and these drives are the last ones at the very end of the chain.

If you mean there are 12 drives using the same cable from the PSU it's likely a problem.

[Pluvichor]

Yes, I bought a few of these

 

https://www.newegg.com/p/N82E16812119915?item=N82E16812119915

 

and connected them to each other to reach every drive. It might actually be more than 12. I forget exactly how I handled the 4 drives at the bottom (in cages)... I know they used a separate line from the PSU because I couldn't make the daisy-chain, er, twist enough to reach them, but I might have managed to include at least 1 in the chain so it might be 13.

 

Is there a limit to how many drives I should connect in this way? Like, should I only use 1 power splitter per PSU cable, or are power splitters a no go entirely?

 

Thank you so much.

[JorgeB]

Very bad idea, don't recommend more than 4/5 HDDs per cable.

[JonathanM]

SATA to SATA power splitters are a no go in my opinion. If you get some 4 pin to SATA power, those can work out ok, probably 4 drives per "molex" type connector would be safe. The SATA power connector has very thin pins and connections, it's not meant for high current.

[Pluvichor]

Hello, just wanted to give an update. I switched the power cables so that the longest daisy chain was 5 disks and when the array started, it seems to be working perfectly. I did a New Config just to be safe, but parity has been building with no errors so far (before many thousand would appear within a few minutes. The speed of the gui seems snappy and responsive instead of laggy (although I haven't done much with it). I need to leave my home until tomorrow so I won't know for sure until then, but it does seem like the issue was using a long power split chain. I'll update tomorrow when I get back tomorrow (hopefully to a finished parity), thank you for all your help.

 

Edit/Update: a day and a a half later and parity is finished building and everything seems to be working great. I have an issue where I can't download the communty applications plugin, but I think that is entirely unrelated to this. Thank you again.

Edited July 16, 2022 by Pluvichor
update