September 15, 20187 yr So, not touching this damn array this time... I have had: Parity + 5 Data Disks for a while... I just pre-cleared 2 disks (one fine, added to array, the other not getting past the unraid signature step onto the post-read, left it alone). I now have Parity + 6 Data Disks. I ran: nohup cp -r /mnt/disk5 /mnt/disk6 & It started and then failed with a "segmentation fault" ... I look to the Main page and see disk missing (disabled, contents emulated) but there's nothing on it yet. I've rebooted a couple of times, all of the disks have been detected (i.e. show up in lsblk) but now I've got a missing disk and a disabled disk I'm suspecting a slightly dodgy PCIe to SATA card but can't be sure. Just don't want to screw up my array for obvious reasons (not twice in a couple of weeks 😧)... --------- I've attached a diag bundle, one where disk6 was disabled (after the attempt to cp data to it) that's 0843 And another one from now, where I've got a missing disk4 for some damn reason - Just checking what port it's plugged into RIGHT. 0941 So the crappy marvel 4 port has 3 HDDs: (serial ending) * E067 * C2L1 * JDJ3 That last one is fine.. But the first 2 connected to that card have gone tits up! raptor-diagnostics-20180915-0843.zip raptor-diagnostics-20180915-0941.zip
September 15, 20187 yr Author Considering buying this: https://smile.amazon.co.uk/10Gtek-Internal-Express-Controller-SAS2008/dp/B01M2AC40Y/ref=sr_1_1?ie=UTF8&qid=1537007716&sr=8-1&keywords=SAS+controller+HBA And flashing it to IT mode: ftp://ftp.supermicro.nl/driver/SAS/LSI/2008/IR_IT/Firmware/IT/ (Use supermicro's FTP site to flash the LSI 3008 to IT mode all the time at work) ------ I've got 4hrs left to get this on Prime tomorrow (Sunday 16th) along with a SAS to SATA breakout cable... I know the disks are fine, they're brand new and underwent a preclear. It's not impossible for it to have failed, sure... But I'm strongly leaning towards the cheapo SATA card (https://www.scan.co.uk/products/lycom-pe-120-ahci-sata-iii-6gbps-internal-4-port-low-profile-pcie-20-host-adapter) Based on this; the 3 disks connected to it had 'errors' on the Unraid Main tab the second that I began to `cp` from disk5 to disk6 (errors appeared on disk4/5/6) and now here I am, a vanished disk4 & a disabled disk6! I don't want to start mixing drives around... I have screenshots of the assignments so that's ok. There's no spare ports, that card has been running with 1 disk on it for a over a year - seems that the moment I add a 2nd and then 3rd drive that it's showing 'weakness' great... I only have 6 onboard SATA + the 4 from this card. I've got no other HBAs or SATA cards, etc. etc. Hence why I'm considering the LSI card so I can hook all the HDDs up to a (hopefully much more solid) card and get the array started again...
September 15, 20187 yr Community Expert Looks like a controller problem, if you can get an LSI to replace the Marvell controller do it.
September 15, 20187 yr Author Oh, I was on the right lines? Excellent, I shall do so (yay more money to spend 😅) Cheers Johnnie! Fingers crossed! (Why me 😂😭)
September 16, 20187 yr Author @johnnie.black Hey LSI card arrived, installed, flashed to IT mode, everything detected again... However I still get errors on Disk6 as soon as I stick it into the array 🤔 It passed a single preclear run with no errors but it hates being in the array - I'm doing another preclear but wanted to start my array and make it safe! So... I've set "No device" for disk6 - the contents will be emulated and thus if I experienced a REAL disk failure whilst in that state - I'd be screwed (right?) I had no data on disk6 so I'm not fussed about emulating an empty 3TB drive Am I correct in thinking I can take a screenshot now of the assignments, go to tools > New config and 'restart' with parity + disk1-5 and be in the green again? And let this preclear run in the background before I think about adding disk6 again? Cheers!!
September 16, 20187 yr Community Expert You can do a new config but parity needs to be synced, only after that the array will be in a protected state.
September 17, 20187 yr Author Turns out it was two DOA WD's, after I ripped them out of the array and started prodding at them some more via Preclear & even Windows they started to show signs of badness! 😂 WD agreed and are shipping some brand new drives out to replace these, yay... Guess I can't close a thread here But 'problem solved' ... I'm incredibly unlucky and now I've got an LSI HBA card, win win I suppose 🤷♂️
November 9, 20205 yr Hei guys !! The monitor cable kind a blown today. The UPS shut down immediately. After starting with the unraid am missing 2 disks now. Not sure what to do. Will it be the smps or the sata cable ? I have an HBA connected to the config. Not sure what and all fried out. Help please !! tower-diagnostics-20201109-1242.zip
November 9, 20205 yr Community Expert Swap cables with another disk and see if the issue follows the cables or the disk.
Archived
This topic is now archived and is closed to further replies.