May 26, 200719 yr I awoke this morning to a locked up unraid (left some automated heavy video processing running against one share last night). The management utility was active and showed 121 errors on the parity drive. I restarted and after remounting all of the drivers (it took 5 minutes for some of the drives as it seemed to be rolling forward through one of the reiser journals, and after about 15 mins it has started a parity check and the shares are visible. Should I be concerned? What happened? Thanks in advance for any advise from the experts. Attached are the relevant portions of syslog overnight:
May 26, 200719 yr Author Well I just had it lock up again at about 53% parity (which took about 6 hours to get that far). Tha management utility showed 100 errors on the parity drive (500GB Seagate SATA). I was unable to get a syslog. I noticed the parity drive had a 3 foot Sata cable (don't know where I got that from), so I replaced it with a more standard 18 inch Sata cable and I'll try checking it again and watching closely at that 53% level. Any advice? Is there a way I can check the parity drive that's quicker than a full parity check? Am I likely looking at a bad/dying sata parity dirve (it's only about 4 months old).
May 27, 200719 yr Author Well, my parity check completed just fine with the shorter cable (finding 112 parity errors). I don't know what to think at this point. I guess I'll keep the drive in the array, but I'll keep it on a short leash for a week (checking status every day, recalculating parity every couple of days).
May 30, 200719 yr Yes keep an eye on the parity drive. A 3 ft cable is a suspect, though spec says up to 1 meter is ok. If you don't mind, what is the nature of your video processing? I'm curious about the I/O pattern and how much file open/close activity might be taking place.
May 31, 200719 yr Author My video processing involved converting every HD video file on a disk (and set of disks) from .TS (MPeg Transport stream) to .MPG (MPeg Program Stream) in bulk 1 at a time unsing VideoReDo's "QuickStream Fix" driven by a Windows Scripting Host automation scipt I whipped together. I doing this for approximately 1.5TB across 5 disks in all. VideoReDo was reading the input files and simultaneously writing the output to the same folder - Typically a 10-18GB file). I imagine it was only 1 file open (handle remained open) for the read and another for the write. Each file takes about 45 mins to process. I was converting the files for Xbox 360 Media Center compatibility with Vista Media Center (Vista MCE no longer streams TS to the 360, even though MCE2005 was fine - *^&() Microsoft ) I may have been doing some smaller tasks on other didk on the array at the same time (download, file mgmt) and at times a parity scan was running as well. (Yeah, I know, take it easy )
Archived
This topic is now archived and is closed to further replies.