February 4, 201115 yr Formulating a plan to recover. Backstory: Gigabyte MB added HPA to one of my drives (aka drive A) when I upgraded from 4.6.5 to 4.7. I tried to use hdparm to fix it, but ended up doing it to my parity drive instead of the drive with HPA, oops. hdparm doesn't work correctly on my parity drive, so I have to use seatools or something else to fix that now. 1) Get new MB without HPA 2) Disconnect parity drive and hook up to other PC, run seatools and set sectors back to correct size 3) Replace MB and reinstall parity drive 4) Boot, verify contents of parity is in place and valid 5) Run hdparm / seatools on A 6) Allow unraid to rebuild the contents of A My only concern is that the contents of the parity are currently wiped because of hpdarm, and using seatools will not allow them to be recovered. If this is the case, I think I have to backup the contents of A to other drives elsewhere, then rebuild parity without drive A and copy the contents back over. Would this even work? syslog_1.txt
February 4, 201115 yr Setting the HPA on the parity drive, then resetting it probably did not affect the data in those last few meg of the disk since it was not written to. Once you reset the parity disk to use to the full size of the disk I think you'll be able to use it to re-construct the other disk. You plan to save the contents of drive A elsewhere won't hurt. as long as you have the time. Then you can re-construct it from the parity drive. Joe L.
February 15, 201115 yr Author So I... Installed my new MB Removed the parity disk and ran seatools for DOS on it to set it to the maximum size, success Removed the other disk and ran seatools for DOS on it to set it to the maximum size, success Reinstalled both drives Booted unraid Saw the array came up successfully, however disk1 had been 'upgraded' Started the rebuild *This is where I get stupid* Mid rebuild I decide to export the drive shares because I want to access some stuff. The server is no longer accessible via http. Not sure if it is continuing the rebuild or if I botched it. Advice? I mean besides the obvious don't be stupid?
February 15, 201115 yr Author It looked like the rebuild was going to take about 16 hours so I am leaving it as is for approximately 24 in case it takes longer. At that point I plan on rebooting and seeing where it is. Please let me know if I should wait longer than that or do something differently. I cannot telnet, ping, browse via smb, or login directly to the server.
February 15, 201115 yr Author Drives are sill hot this morning so they are spinning and I assume the recovery is proceeding successfully. I will wait until they have spun down to reboot.
February 16, 201115 yr Author Hmmm....still hot and spinning. How long does a 2GB drive rebuild normally take when both the Parity and drive are on the MB?
February 16, 201115 yr Hmmm....still hot and spinning. How long does a 2GB drive rebuild normally take when both the Parity and drive are on the MB? About the same time as an initial parity calc.
February 17, 201115 yr Author After approx 56 hours the drives were both still spinning, but I assumed it had failed and I went ahead and reset via the power button. When it booted back up it automatically began the rebuild, I watched it get to 0.6% and then left it for a couple hours (not off unraid). When I came back and hit refresh the connection timed out. I can not reach the server or login directly to the console either. Looks to be locked up.
February 18, 201115 yr Author That's what I meant by tried to login via console. It would not let me type a username. Completely locked. I think it has to do with something trying to connect to samba. I have tried 3 more times and it makes it somewhere from .3% to 1% and then locks up. I am trying again but I unplugged the network cable this time. I will continually check if it has locked up by typing on the actual server and let it stay disconnected for maybe 24 hours before plugging in the cable. Is there a way to monitor the rebuild process from the actual server?
February 18, 201115 yr Author I came back and checked on it several times. After about 25 minutes the monitor turned to this: I am guessing my parity drive was invalid because of the HPA stuff, and that's why the rebuild is failing.
February 18, 201115 yr Author Under power, BIOS has ACPI Suspend Type with 3 options: S1&S3, S1, S3. It is currently set to S1&S3. Should I change it?
February 18, 201115 yr Author Changed to S1, still locked up, and did it 60 seconds after I logged in so probably not a sleep problem I am guessing.
February 18, 201115 yr I came back and checked on it several times. After about 25 minutes the monitor turned to this: I am guessing my parity drive was invalid because of the HPA stuff, and that's why the rebuild is failing. I cannot see a way an HPA can cause that type of error on a monitor. That is video RAM that has been corrupted somehow. That is far more likely to be power-supply/memory/motherboard/cpu related.
February 18, 201115 yr Author Maybe that video corruption is related to S3. It seems that when I use S1 only the corruption does not appear. It still freezes however. It just locks up at the command prompt and eventually goes black. I rebooted into S1 and am running top, just to give it something to do repeatedly so it doesn't go to sleep. Let's see what that does.
February 18, 201115 yr Author And again, it locked up. I think my next step is to boot and attempt to stop the recovery, and see if unraid still freezes.
February 18, 201115 yr Author So I did some extra stuff to try to see what was locking it up. I plugged it back into the network. I booted and immediately cancelled the rebuild and stopped the array. I left it running for 30 min and it did not freeze. I then started the array and again cancelled the rebuild. I left it for 30 minutes and it did not freeze. I was able to navigate and browse the shares from another computer. It allowed me to browse. This is where it did something unexpected. I was able to browse to disk1 share (the disk I have been unsuccessfully rebuilding), and it still appears to have all my files on it. syslog is attached. Version is 4.7. Will run memtest now. syslog.txt
February 18, 201115 yr Author Locked up in middle of memtest. Did not show failure message, just got stuck at 24% overall, 99% current test status. Reseated memory, running again. This is the same memory I was using, just a different MB.
February 18, 201115 yr Locked up in middle of memtest. Did not show failure message, just got stuck at 24% overall, 99% current test status. Reseated memory, running again. This is the same memory I was using, just a different MB. If you have more than one stick, try each one separately, and if they pass try them together. It sounds like any time you put stress on the system is fubar's itself. I would be looking at the motherboard first in this instance.
February 18, 201115 yr Author Ran individually and both sticks passed in slot A-1. Previously they had failed in dual channel configuration (A1, B1). Put them both back in in non-dual channel configuration (A1, A2), and running again. Maybe it's that B2 slot.
February 18, 201115 yr Author Success as non-dual channel. In the meantime I pulled disk 1 and am coping contents off on another pc in case I lose it.
February 18, 201115 yr Author Passed in non-dual channel configuration. Put sticks back into dual channel configuration, but swapped the sticks as compared to original configuration. Passed memtest twice in a row.
Archived
This topic is now archived and is closed to further replies.