July 19, 20187 yr Community Expert You need to find what's the problem first, i.e., try a different power supply, HBA, cables, etc, trying to repair the drives or array like this is likely to damage the data even more.
July 19, 20187 yr 1 hour ago, Maticks said: but its 140Watts on an 850Watt power supply its hardly stressed How hard you push a power supply doesn't always correspond to failure mode. You need to try a different power supply and remove all your current power leads to verify whether or not it's the issue. Your symptoms very much sound like a bad PSU. DO NOT TRY TO REUSE MODULAR CABLES (unless you know how and why to use a DVM to verify functionality)
July 19, 20187 yr Author i think it was too many drives on a sata power cable. the disks were stalling. i had 8 fans hanging off that sata power lead as well it's clear that was the issue. But i can't work out how to repair this there is no file system repair
July 19, 20187 yr Community Expert 57 minutes ago, Maticks said: no making a parity disk here it seems ul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079200 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079200 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079200 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079200 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079208 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079208 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079208 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079208 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079216 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079216 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079216 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079216 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079224 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079224 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079224 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079224 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079232 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079232 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079232 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079232 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079240 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079240 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079240 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079240 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079248 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079248 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079248 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079248 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079256 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079256 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079256 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079256 Jul 19 22:21:42 Vault kernel: md: disk1 read error, sector=1034079264 Jul 19 22:21:42 Vault kernel: md: disk4 read error, sector=1034079264 Jul 19 22:21:42 Vault kernel: md: disk5 read error, sector=1034079264 Jul 19 22:21:42 Vault kernel: md: disk9 read error, sector=1034079264 These are not filesystem errors.
July 19, 20187 yr Author looks like the clicking was the drives powering down every few seconds for 24 hours.. i guess this would break writes and reads. sounds like blow it away and restore from backup. i've distributed the power now between the sata drives but having those 140mm fans on the same power lead as the sata drives sucked a lot of amps it seems.
July 19, 20187 yr Author i know what it is now. All the drives with read errors are on my LSI card all the drives that are fine in the array are on my onboard SATA ports. My LSI card must be having some kind of failure Bingo... Reports online when an LSI card is about to fail two drive start getting read errors then all drives start having issues on the controller. Doing a parity rebuild since I blew away my parity drive and have most of my data in tact need to order a new LSI card. Let's see if it finishes in the morning. Edited July 19, 20187 yr by Maticks
July 19, 20187 yr 1 hour ago, Maticks said: i know what it is now. All the drives with read errors are on my LSI card all the drives that are fine in the array are on my onboard SATA ports. My LSI card must be having some kind of failure Bingo... Reports online when an LSI card is about to fail two drive start getting read errors then all drives start having issues on the controller. Doing a parity rebuild since I blew away my parity drive and have most of my data in tact need to order a new LSI card. Let's see if it finishes in the morning. You sure the card isn't just overheating?
July 19, 20187 yr Author Parity rebuilt error free. It's the LSI card for sure I even included disk 1 which was full of errors before, 0 when it's on the onboard SATA. The LSI card is barely warm to the touch it has good airflow over the card and a heatsink. Tried a different PCI slot as well. Few forum posts I've looked for describe my issue exactly two drives start erroring then everything starts going wrong. Googling lsi card two drives read errors, every post is exactly my two days of hell. The good news is my data drives are all good I just need a new LSI card. I am certainly going to buy another and have a sas cable in each, looks pretty easy to resolve if it happens again. I just decided to only include the good data drives in my new config with the parity drive that was error free. When I looked at the case during the rebuild the LSI card read and write lights weren't flashing and I realised that was because all my drives were the onboard controller on the motherboard. Powered down reshuffled some sata power and surprise the bad data LSI ports drives were all fine now. Could still be the cables not the card I guess, but I only touched one cable and both failed in the end. The LSI card toward the end after a system reboot would actually only detect the drive during the raid main screen selection for about 5 seconds then it would come up no device in the array popup. The card would be around 3 years old. Edited July 19, 20187 yr by Maticks
Archived
This topic is now archived and is closed to further replies.