June 27, 200917 yr Hello, I think my unraid server was still on when there was a power cut yesterday. Either way, now when I turn the server on the array doesn't start and on the unraid page all drives have a blue light next to them, which I believe means new disk. What steps should I take to check that my data is still safe and get the array working again? Thanks, bbkarn
June 27, 200917 yr Hello, I think my unraid server was still on when there was a power cut yesterday. Either way, now when I turn the server on the array doesn't start and on the unraid page all drives have a blue light next to them, which I believe means new disk. What steps should I take to check that my data is still safe and get the array working again? Thanks, bbkarn Post a syslog. It is the only way to know what is happening. Instructions under "Troubleshooting" in the wiki. Do not press "Restore" Do not remove or replace drives, do not change wiring... The content of the syslog will provide clues as to how to best proceed. We will advise once we see it. Joe L.
June 27, 200917 yr Author Thanks for the info, syslog attached to this post. I have not tried anything beyond a reboot. bbkarn
June 27, 200917 yr The lines in the syslog that indicate what happened are these: [pre] Jun 27 14:05:53 Media kernel: md: invalid superblock checksum Jun 27 14:05:53 Media kernel: md: initializing superblock [/pre] It indicates the checksum on the contents in the super.dat file on your flash drive do not agree with the contents. You are a prime candidate for the "Trust my Parity" procedure as described in the wiki. It will force a rebuild of the super.dat file and then it will set about verifying that your parity is correct. (It might find a few errors, if something was not written to a disk before the power failure, but it will correct them when it does) when you start the array it will start a full parity check. Let it complete, do not cancel it. The procedure is described here: http://lime-technology.com/wiki/index.php/Make_unRAID_Trust_the_Parity_Drive,_Avoid_Rebuilding_Parity_Unnecessarily Oh yes... think about getting a UPS... Joe L.
June 27, 200917 yr Author Many thanks, the procedure you recommended has brought everything back with the parity check now running. And yes, a UPS suddenly seems like quite a sensible thing to have... bbkarn
August 25, 200916 yr I just had the same problem. We have had many power failures but this is the first time this has happened. I have atteched my syslog and would really appreiciate it if some one could take and look and see if I need to do the same as the previous post. I do not have the restore option though, it is grayed out. Only option is start. Thank You
August 25, 200916 yr I just had the same problem. We have had many power failures but this is the first time this has happened. I have atteched my syslog and would really appreiciate it if some one could take and look and see if I need to do the same as the previous post. I do not have the restore option though, it is grayed out. Only option is start. Thank You You have the same option... and the same problem From your syslog: [pre]Aug 24 20:04:40 Server kernel: md: invalid superblock checksum Aug 24 20:04:40 Server kernel: md: warning! initializing superblock[/pre] Below the "restore" button is a checkbox. Checking it will enable the restore button. Beware, "restore" does NOT restore data...NEVER use it if you have a disk failure and you are expecting to rebuild the data from parity, as it immediately throws away the parity data. It is mis-labeled (in my opinion) as it is a "Set Initial Configuration" button... It sets the config to the currently assigned and working drives and then starts a parity calculation on those drives. Any prior parity calculations are thrown away unless the "trust-my-parity" process is followed. In my opinion, you are a prime candidate for that procedure. And also a prime candidate for a UPS. Joe L.
August 25, 200916 yr I will follow the instruction in this link http://lime-technology.com/wiki/index.php/Make_unRAID_Trust_the_Parity_Drive,_Avoid_Rebuilding_Parity_Unnecessarily to the letter then. As for UPS, yes I agree. The problem I have is that we have had about 3 different UPS's at work and every one of them when you really finally need them didn't work.
August 25, 200916 yr As for UPS, yes I agree. The problem I have is that we have had about 3 different UPS's at work and every one of them when you really finally need them didn't work. I know... the batteries only last about 3 years or so... You don't need a huge UPS for the unRAID server. I have one that was on sale for about $39. on my server. It gives about 7 minutes run time, but it is monitored by the unRAID server, so when power is lost, and only a few minutes of run time is left, it stops the server and cleanly powers down, including itself.. waiting for power to return. It is an APC brand. It does not shut down itself or the server for short outages. Instead, I just see a note in the syslog that power was lost momentarrily. Like this, from a week or so ago Aug 16 12:09:55 Tower apcupsd[2073]: Power failure. Aug 16 12:09:57 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 18 20:17:19 Tower apcupsd[2073]: Power failure. Aug 18 20:17:21 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 19 20:22:14 Tower apcupsd[2073]: Power failure. Aug 19 20:22:16 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 20 12:37:37 Tower apcupsd[2073]: Power failure. Aug 20 12:37:38 Tower apcupsd[2073]: Power is back. UPS running on mains. and here Aug 21 23:02:01 Tower apcupsd[2073]: Power failure. Aug 21 23:02:02 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 22 14:11:06 Tower apcupsd[2073]: Power failure. Aug 22 14:11:07 Tower apcupsd[2073]: Power is back. UPS running on mains. and here Aug 23 19:26:56 Tower apcupsd[2073]: Power failure. Aug 23 19:26:58 Tower apcupsd[2073]: Power is back. UPS running on mains. Aug 23 19:36:34 Tower apcupsd[2073]: Power failure. Aug 23 19:36:35 Tower apcupsd[2073]: Power is back. UPS running on mains. Now you see why I have a UPS... Summer lightning storm season here in North Carolina, and lots of tiny power blips. ( 8 in 10 days ) Joe L.
August 25, 200916 yr As for UPS, yes I agree. The problem I have is that we have had about 3 different UPS's at work and every one of them when you really finally need them didn't work. I know... the batteries only last about 3 years or so... You don't need a huge UPS for the unRAID server. I have one that was on sale for about $39. on my server. It gives about 7 minutes run time, but it is monitored by the unRAID server, so when power is lost, and only a few minutes of run time is left, it stops the server and cleanly powers down, including itself.. waiting for power to return. It is an APC brand. It does not shut down itself or the server for short outages. Instead, I just see a note in the syslog that power was lost momentarrily. Like this, from a week or so ago Aug 16 12:09:55 Tower apcupsd[2073]: Power failure. Aug 16 12:09:57 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 18 20:17:19 Tower apcupsd[2073]: Power failure. Aug 18 20:17:21 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 19 20:22:14 Tower apcupsd[2073]: Power failure. Aug 19 20:22:16 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 20 12:37:37 Tower apcupsd[2073]: Power failure. Aug 20 12:37:38 Tower apcupsd[2073]: Power is back. UPS running on mains. and here Aug 21 23:02:01 Tower apcupsd[2073]: Power failure. Aug 21 23:02:02 Tower apcupsd[2073]: Power is back. UPS running on mains. and this Aug 22 14:11:06 Tower apcupsd[2073]: Power failure. Aug 22 14:11:07 Tower apcupsd[2073]: Power is back. UPS running on mains. and here Aug 23 19:26:56 Tower apcupsd[2073]: Power failure. Aug 23 19:26:58 Tower apcupsd[2073]: Power is back. UPS running on mains. Aug 23 19:36:34 Tower apcupsd[2073]: Power failure. Aug 23 19:36:35 Tower apcupsd[2073]: Power is back. UPS running on mains. Now you see why I have a UPS... Summer lightning storm season here in North Carolina, and lots of tiny power blips. ( 8 in 10 days ) Joe L. Can you give me the model of the UPS? any problem setting it up? Is there a guide to do so? I have few APC UPS and few Belkin UPS. Basically every workstations in my house has a UPS but having one that works tightly with the NAS is very blessing.
August 25, 200916 yr Can you give me the model of the UPS? Most any APC will work. I have a APC Back-UPS ES 750any problem setting it up?No, you do need to install a "powerdown" package too. It shuts the unRAID array down cleanly. Is there a guide to do so? Look here in the wiki: http://lime-technology.com/wiki/index.php/UnRAID_Topical_Index#UPS_support Absolutely easiest is to install the unMENU add-on, then use its package manager page to download and install and configure the "apcupsd" UPS software and the powerdown script. (It is configured for a USB connection to the UPS, but you can see how to edit the config file if you have a serial port UPS connection.) Joe L.
August 25, 200916 yr I will have to look into a UPS sounds like a good deal. I just tried to get into the server through its web adress and I can't get in. Can't get in through TELNET either. But I can ping it and that comes back...... I haven't tuched it since last night when I got the syslog to post here. Guess I will have to shut it off and restart it later (at work now).
August 26, 200916 yr Ok so I went through the instructions and and it all started back up and started doing its parity check, started about 6:00 last night. At 10:00 I checked it and it was about 20% done I believe. Now thismorning I went to pull up the status in explorer and I can't get in again. Can't acess the disc's from windows, I can ping it and it come back. So I tried Telnet and got the login prompt, but interupted and had to close the window, tried again 5 minutes later and now I can't Telnet it. I don't know what is going on....... Now I don't know if it finished its parity check before it crashed or what ever it did and I can't get at the syslog?
August 26, 200916 yr Ok so I went through the instructions and and it all started back up and started doing its parity check, started about 6:00 last night. At 10:00 I checked it and it was about 20% done I believe. Now thismorning I went to pull up the status in explorer and I can't get in again. Can't acess the disc's from windows, I can ping it and it come back. So I tried Telnet and got the login prompt, but interupted and had to close the window, tried again 5 minutes later and now I can't Telnet it. I don't know what is going on....... Now I don't know if it finished its parity check before it crashed or what ever it did and I can't get at the syslog? You syslog has probably filled all available memory, and the linux out-of-memory process is killing processes in an attempt to gain more free memory. The errors are probably as a result of a disk failure of some kind. But you won't know for sure until you can capture them. Go through the same process of re-starting the server once more, but this time, either telnet to it and run tail -f /var/log/syslog or log in at the system console and run tail -f /var/log/syslog This time, you'll see the errors and can report on what is happening. Joe L.
August 26, 200916 yr It is strange, very intermitent. I can get in for a short time, I did manage to get in long enough to create the syslog. But now I can't get the syslog off the flash. I got in far enough to see that it is on the flash, just couldn't copy it.
August 26, 200916 yr If you copied the syslog file to the flash drive, you can just reboot... Once back up it should be visible from the LAN.
August 26, 200916 yr If you copied the syslog file to the flash drive, you can just reboot... Once back up it should be visible from the LAN. Yea I have to do that when I get home. Trying to do it from work over a VPN now so I can't push the button.
August 26, 200916 yr If you copied the syslog file to the flash drive, you can just reboot... Once back up it should be visible from the LAN. Yea I have to do that when I get home. Trying to do it from work over a VPN now so I can't push the button. If you can log in long enough to type reboot it should do it. It will not stop cleanly, and will want to do a parity check when it re-starts, but I think it would need that anyway if you were to push the power button.
August 26, 200916 yr I finally got the connection to stay long enough to get one syslog off from it. Here it is.
August 26, 200916 yr I finally got the connection to stay long enough to get one syslog off from it. Here it is. I don't see anything really odd, other than the superblock file is still not readable. There is no evidence of the massive errors I was expecting filling the syslog. Joe L.
September 3, 200916 yr According to the syslog, you appear to be connecting from a different subnet, which may be complicating the connection. That is unusual to me, but Joe will know better if that is what is causing the inconsistent connections. Your startup is rather unusual too, as it looks like you are using a kicker disk(?), or some alternative way of booting your server. The following is your boot command line: Aug 25 16:36:36 Server kernel: Kernel command line: root=/dev/sda1 Can you give us some detail about how you boot?
September 3, 200916 yr I do use a kicker disc yes, my mother board will not boot from usb so this was my way around it. The server is at my house and I connect to it from work through a VPN connection using Sonic Walls home is on a 192.168.77..... and work is 192.168.76...... that may be what you are seeing. I did get everything back up and going again and it works just fine at home. But for some strange reason I can't get on from work anymore. I use to go into the server all the time from work for the last year. Nothing has changed in our VPN setup or on the server so I don't know what is going on. I can access other printers and computers at home fine, it is just the UnRaid box that is being a problem.
Archived
This topic is now archived and is closed to further replies.