May 11, 201412 yr Hello everyone, I'm getting parity incorrect issues everytime I run a parity check. Im running UNRAID 6.0 Beta 5 Things I have tried/verified: 1. Checked the smart status of all drives and they all seem okay 2. No red warnings in the webui and no write errors 3. Let multiple parity checks run until completion in the following order: - Ran one (errors) - Ran one the next day (errors) - Ran one a week after (errors) 4. Ran initconfig and rebuilt the parity drive - Ran a parity check (errors) Its not the same locations each time throwing the errors and Im not sure what else it could be other than a failing disk, which SMART indicators seem to think its now. I've verified no errors of any note in the syslog either. Below is a small part of the parity errors. mdcmd (44): check CORRECT May 11 17:23:23 Media kernel: md: recovery thread woken up ... May 11 17:23:23 Media kernel: md: recovery thread checking parity... May 11 17:23:23 Media kernel: md: using 1536k window, over a total of 1953514552 blocks. May 11 17:23:28 Media kernel: md: correcting parity, sector=1085040 May 11 17:24:16 Media kernel: mdcmd (45): nocheck Any advice?
May 11, 201412 yr over the years there have been two causes for that type of symptom. 1. Bad or incorrectly configured memory. Run a memory test, through several cycles, preferably overnight, expect no failures. 2. A bad disk drive. It randomly returns incorrect data, but emits no other signs of errors. These are the hardest to find and isolate, since you have no way of knowing which disk it might be. One unRAID user created a test script to repeatedly read, using "dd" the same set of blocks on a given disk and to compute the checksum of those blocks. Assuming you are not writing to the disk, the checksum will not change. If it does, that disk is suspect. Perhaps somebody will remember the thread with that script and point you to it. Joe L.
May 12, 201412 yr Author After running memtest overnight I got a few errors which definitely could be causing issues. I pulled the RAM out and place them, one at a time into slot 0 on the MB and ran memtest again and go no errors. Placing the memory back in both slots causes errors. I also notice a lot of warnings regarding the USB drive and unable to access /boot/config/super.dat Ran chkdsk from a windows machine and it seems that drive is having issues as well.
May 12, 201412 yr What motherboard do you have? Chances are there could be an issue with slots and timing, or just timing in the bios. I had an ABIT AB9 PRO that was notorious for overclocking the ram. If I ever cleared the CMOS, I had to get into the BIOS as fast as possible to relax the ram timings. Heat can also be an issue, does the ram have heatspreaders?
May 12, 201412 yr How long are you running the memtest? I'd suggest 24 hours (more for lots of memory). Don't run a machine with bad memory. It can cause all sorts of problems. Is it possible that your second DIMM slot is obstructed? You might try blowing it out.
May 15, 201412 yr Author You were correct, it was an issue with ram timings. Setting them back to valid setting Parity checks run without errors. Thanks all for the help. What motherboard do you have? Chances are there could be an issue with slots and timing, or just timing in the bios. I had an ABIT AB9 PRO that was notorious for overclocking the ram. If I ever cleared the CMOS, I had to get into the BIOS as fast as possible to relax the ram timings. Heat can also be an issue, does the ram have heatspreaders?
Archived
This topic is now archived and is closed to further replies.