I have an LSI SER523 4-Port Raid card running raid-5 with three WD 250GB disks. This set-up has been running peachy for a few years now, and lately the machine has been mysteriously powering itself off.
I originally assumed it to be a power supply issue, but that's working just fine.
After more or less disassembling the machine and plugging things back in one at time, now the raid card is beeping loudly and shows an error during boot. i can choose to continue and boot my machine without using the drives thru raid, because i have a separate HD as the boot drive.
If i go into the raid management screen, i can see that my single logical drive is offline and it's reporting that all three of my physical drives have failed. this sent up red flags to me that it might just be a bad raid card because my karma can't be so bad as to have all three dives fail at the exact same time. the management console can detect the drives and display information on each one. each drive appears to be spinning, no funny noises, etc. I attempted to rebuild the drives thru management, but for each drive the rebuild just errors out right away. I know that at least part of the raid card works as I can easily silence the alarm through the management screen.
I'm not sure if a failing raid card will result in a system powering itself off, but has anyone seen or experienced anything like this before? i'm stumped.
do you have another identical card to test with? You should be able to read the config off hte drives into NVRAM and boot with the array intact if its the card itself. Make sure you tage the port and drive locations before replacing the card. Do NOT recreate the array or initialize the drives in any way.
If this is not an option there are ways of potentially recovering the data if that is the issue.
Yep, I was going to suggest something in the same vain: divide and conquer (try to narrow down the issue), however I was going to suggest pulling a drive from the array (this presupposes you'd be able to later rebuild the array) and boot into a livecd to test out the drive (liveCD to prevent SW issues from muddling the results), read the SMART status, run some SMART tests, possibly even format the disk and do a pragmatic test, if that all checks out then that would certainly point to the card, in which case you'd need to get an identical replacement and rebuild the array (at least this type of approach worked with my 3Ware Escalade 4 port)