RAID 5
Friday 14th July, 2006 09:48 Comments: 0
I\'ve decided to try recreating one of my RAID 5 arrays as a set of 7 disks, leaving one spare, instead of the original set of 8. The idea is that if the RAID array were to fail because it lost a hard disk (yes, I know, how exactly do you lose a hard disk when it's still plugged in, spinning, and secured in the case with some wires and screws?), the system would rebuild itself on the spare drive. I don\'t know if the system would then beep for the 2.5 hours or so it\'d take to rebuild, or even if it would keep beeping after it's been recreated, but it's got to be a bit safer to have a spare lying dormant - I believe Chris has this setup with his drives. This meant I had to delete the old array, which didn\'t have that much data on it, but I\'d brought my 120GB disk back and left the 200GB one at my parents. I could have waited until I got back on Sunday, after collecting the drive this weekend, but I decided to cram as much as I could onto every spare drive (and a couple of DVDs) and I\'ll just sacrifice the rest. Most of it's old stuff I have on some unhealthy CD-Rs, some are things I can get away with losing. I\'ve only lost about 40GB worth of stuff - probably as big as some people's hard disks are - and the new array is building itself up as I type this.
If it does work, I may take a screwdriver (or is there a jumper on the cards?) to the loudspeakers on the RAID cards and permanently disable them. I\'ve set it to notify me by email, and I\'ve even setup a system where I can get SMS messages sent to my mobile if an error occurs, so it's not like I won\'t know about it if something goes wrong.
The good news is that I\'ve only had errors recently, I haven\'t had the array fail (touch wood), and I\'m wondering if it's because I changed power management to never spin down the hard disks (my theory was if it didn\'t spin up quickly enough the array would drop it). Chris reckons it could also be SMART that is causing issues, but to disable it I\'d have to connect each SATA disk up to access it directly - going through the RAID controller hides the disk from the software that lets me configure SMART - and this would take forever and involve undoing a hideous amount of wires. I\'ve also got several fans in the case to try and keep the disks and controller cards cool, two fans I\'m quite proud of (neatly and clevery attached with a Zalman bracket), the other 3 are just resting, as the case is on its side right now. I\'m still trying to work out what causes the random errors, but at least they\'re just errors, and not the death of the array. They only seem to occur when writing data, which suggests the problem lies with the controller card when it writes the parity data, so if this new smaller RAID array still has issues, I may be tempted to switch to software RAID with Windows, and if that cures the problem I may try and move the other array to software RAID - after obtaining the 200GB drive (and maybe some others) from my parents\' house, so I don\'t lose any data this time!
If it does work, I may take a screwdriver (or is there a jumper on the cards?) to the loudspeakers on the RAID cards and permanently disable them. I\'ve set it to notify me by email, and I\'ve even setup a system where I can get SMS messages sent to my mobile if an error occurs, so it's not like I won\'t know about it if something goes wrong.
The good news is that I\'ve only had errors recently, I haven\'t had the array fail (touch wood), and I\'m wondering if it's because I changed power management to never spin down the hard disks (my theory was if it didn\'t spin up quickly enough the array would drop it). Chris reckons it could also be SMART that is causing issues, but to disable it I\'d have to connect each SATA disk up to access it directly - going through the RAID controller hides the disk from the software that lets me configure SMART - and this would take forever and involve undoing a hideous amount of wires. I\'ve also got several fans in the case to try and keep the disks and controller cards cool, two fans I\'m quite proud of (neatly and clevery attached with a Zalman bracket), the other 3 are just resting, as the case is on its side right now. I\'m still trying to work out what causes the random errors, but at least they\'re just errors, and not the death of the array. They only seem to occur when writing data, which suggests the problem lies with the controller card when it writes the parity data, so if this new smaller RAID array still has issues, I may be tempted to switch to software RAID with Windows, and if that cures the problem I may try and move the other array to software RAID - after obtaining the 200GB drive (and maybe some others) from my parents\' house, so I don\'t lose any data this time!