Raid Repair Now wont boot - Other mounting problems
OK So it has been a long 3 days of just frustration and waiting.
Here is my situation. 1) Server went down and got booted into a rescue mode where I was able to use putty to find out that my md3 (soft raid) was degraded which is funny cause that is just the /var not the / which in theory should still let the server boot i just wouldn't have my www files. 2) I managed to repair the raid using mdadm and status is all good 3) I reboot and the server never comes back online. In normal Mode (boot from HD) So I put it back into the rescue mode 4) I used to be able to go into the rescue mode and type Code:
mount /dev/md3 /mnt/ 5) Now couple reboots later still not booting I can't run the Code:
mount /dev/md3 /mnt/ Code:
root@rescue:/var/log# mount /dev/md3 /mnt Any help would be great and Id even pay for some help. I have AIM/MSN/SKYPE/GTALK if anyone knows this stuff well and can lend a quick hand.. Thanks |
What is the layout of your disks? That is, how are the physical drives partitioned and how are the various md devices configured?
In particular, which drives/partitions are part of md3? What exactly did you do in step #2? Have you checked the S.M.A.R.T. status of your drives with smartctl? |
Quote:
I have 2 disks. (this is an OVH server but) I have them laid out as such sda ---- sda1 = Raid 1 40GB / ---- sda2 = swap ---- sda3 = Raid 1 ~ 1.7TB /var sdb ---- sdb1 = Raid 1 40GB / ---- sdb2 = swap ---- sdb3 = Raid 1 ~ 1.7TB /var Therefore the Raids are md3 = /var md1 = / md1 and md3 with the corresponding sda1 sdb1 , sda3,sdb3. Here is my fdisk Code:
root@rescue:~# fdisk -l Code:
root@rescue:~# fsck -fc /dev/sda1 Code:
root@rescue:~# fsck -fc /dev/md1 |
You still haven't told us what you did in step #2.
What does mdadm --misc --detail /dev/md3 say? |
Quote:
so i ended up doing this Code:
mdadm /dev/md3 --manage --add /dev/sdb3 mdadm --misc --detail /dev/md3 Result... BOTH md1 (boot) and md3 are clean Code:
root@rescue:~# mdadm --misc --detail /dev/md3 |
Quote:
That means the RAID 1 array was degraded, not broken. You should still have been able to mount /dev/md3. The fact that you couldn't, indicates that the data on /dev/sda3 is corrupt. You then added /dev/sdb3 to the degraded array with mdadm /dev/md3 --manage --add /dev/sdb3. Wouldn't that initiate a synchronization, causing the entire /dev/sdb3 to be overwritten with the (known corrupt) data from /dev/sda3? Could you post the output from smartctl -a /dev/sda? |
Quote:
OUTPUT: this is for A Code:
root@rescue:~# smartctl -a /dev/sda Code:
root@rescue:~# smartctl -a /dev/sdb If I read this correctly the B drive appears to be failing .. could this cause it not to boot? Even though the A drive is just fine? Even if that is so i should be able to Mount just A and look at it but i can't do that. |
Maybe the mdadm terminology is a bit off, because "degraded" is one possible state of a RAID array, not the state of any single member of an array.
After you added /dev/sdb3 to /dev/md3, what (if any) other mdadm commands did you run? Anyway, /dev/sda seems good from the S.M.A.R.T. data. Specifically, "Reallocated_Sector_Count" and "Current_Pending_Sector" are both 0. Could you post the same data for /dev/sdb? |
Quote:
I just updated my post with the B data below the A data.. That drive appears to be bad.. as for what else i ran i just used this Code:
cat /proc/mdstat If you look here http://help.ovh.co.uk/RaidSoft i might have managed to mess up the very bottom of that cause the swap commands did nothing and error ed. Might have not put the right letters idk im stuck. I know my data is good on drive a just not accessible for some reason :( THIS IS JUST A SAMPLE Code:
Personalities : linear raid0 raid1 raid5 |
That is one seriously broken drive. You should unplug /dev/sdb immediately, or at the very least use mdadm /dev/md3 --manage --fail /dev/sdb3 (and repeat the command for md1 and /dev/sdb1).
|
Quote:
I am not able to access the server psychically as it is in another country lol will these commands do the same thing as unplugging the drive and letting it as a single drive server until the Datacenter is able to put a new drive in? Code:
mdadm /dev/md3 --manage --fail /dev/sdb3 |
Alright i did both of those commands and got this output
Code:
root@rescue:~# mdadm /dev/md1 --manage --fail /dev/sdb1 Should I tell the server to Boot from the Hard Drive Now??? Or do i need to change other things to get the server to boot.. shouldnt the raid just say hey there is a good drive here we can use this...? |
Yes, that should do the trick.
|
Ok here goes nothing on the reboot
|
Well system never comes online after reboot :( Something must be wrong... should i be able to mount anything?? I just like to rsync the data to somewhere and start over
|
All times are GMT -5. The time now is 08:18 AM. |