March 23, 201313 yr So, my unraid server is acting very strange after a power outage (momentary interruption and restoration of power) First I noticed that it was not showing up on the network I connected a monitor & keyboard and booted. Unraid seemed to boot normally but was then unresponsive to keyboard commands at the command line. I re-booted and was able to enter BIOS and navigate normally (no delay in keystrokes). I re-booted again and was able to navigate at the initial limeware unraid/memtest menu (keystrokes normal). Selected unraid and for for just a couple of seconds I was able to type at the command line... enough to type "ro" and then it appears to stop accepting commands from the keyboard. Although, about a minute later it completes my "root" and enter keystroke. It seems to carry out the keystrokes minutes after they are typed so it looks like it's not accepting them at all. Also not showing up on the network. No response when I ping it, not showing up in windows explorer, etc... ifconfig eth0 shows - error fetching interface information: device not found My best guess is that the installation is hosed and I should just reformat my flash drive and hope for the best, is that right? Any suggestions before I go that route? If I do reinstall Unraid will I have to start preclear and everything else all over again? It took me a week to setup 6 drives and I don't have anywhere to put my data while I preclear... Help...
March 23, 201313 yr Author Or , I could just replace the bzroot and bzimage files from a clean download and that would essentially replace the OS itself w/o my settings... Right?
March 23, 201313 yr Author Argh, I just replaced bzroot, bzimage, syslinux.cfg & syslinux with fresh files from a new download. Left all my other files on my flash drive... Is there another file I should have replaced or would I be best off with a full reformat and reinstall? If I start over I can save my config director, right? Should I also save my preclear_reports directory? Considering what i've replaced is it starting to look more like a hardware issue? FYI - BIOS is seeing all the drives properly...
March 23, 201313 yr Author Ok, this time I reformatted my USB, copied the unraid files over, ran "make_bootable" as admin and then copied over my config directory backup to the fresh install. Same behavior, BIOS ok, limeware menu ok, get to command line and the first couple of keystrokes are accepted then everything comes to a halt... Am I right in thinking it sounds like hardware at this point?
March 23, 201313 yr Sounds more like multiple devices on the same IP address. (or worse, on the same MAC address) Are you using fixed IP addresses? or dynamic? In your config/network.cfg file, is there a HWADDR=XX:XX:XX:XX line, if so, delete it and reboot. Joe L.
March 23, 201313 yr Author This is the contents of my network.cfg file.... # Generated network settings USE_DHCP=yes IPADDR= NETMASK= GATEWAY=
March 23, 201313 yr Author Oh, and I'm sure you could see from the DHCP status but I am using dynamic IPs
March 23, 201313 yr Author Here's that log. Took a while... although unraid seems to respond to my keystrokes it takes minutes for a single command to appear on screen and run. syslog.txt
March 23, 201313 yr Author Is there some sort of process that unraid puts the drive through after an interruption of power? Could the OS be acting appropriately and running through some sort of check that's tying up the processor and causing the lag at the command line?
March 23, 201313 yr Author So, after some more research I ran the top command and found that 41% of my CPU is tied up with a command called mdrecoveryd. Then I have 6 processes running called kworker/x:y where x is either one or two and y is 1, 2 or 3. Each of those processes is at 6% cpu... That's not a maxed out processor so I'm not sure if it can explain the command line lag but it seems abnormal, right? is mdrecoveryd something that usually runs after an unexpected shutdown. Does this explain anything?
March 23, 201313 yr So, after some more research I ran the top command and found that 41% of my CPU is tied up with a command called mdrecoveryd. Then I have 6 processes running called kworker/x:y where x is either one or two and y is 1, 2 or 3. Each of those processes is at 6% cpu... That's not a maxed out processor so I'm not sure if it can explain the command line lag but it seems abnormal, right? is mdrecoveryd something that usually runs after an unexpected shutdown. Does this explain anything? Well, 100% cpu does normally make things take longer, due to the fact they're added to the top of the stack and take awhile to hit the bottom to be processed. Anyway:- https://www.linuxquestions.org/questions/linux-general-1/mdrecoveryd-what-is-this-145810/ "It's the recovery daemon for raid arrays." I'd leave it running until someone explains what it does in more detail, even more so because it's mdrecoveryd.
March 24, 201313 yr Author So, I let it sit overnight and now on the top command screen I seen that mdrecoveryd has scaled back it's CPU suage and now the command unraidd is on top of the list with roughly 60% usage... AND the command line is reponsive to keystrokes now. I still don't see my server in my Windows network and ifconfig eth0 gets me "eht0 : error fetching information : Device not found" I restarted the machine and found that the command line was still responsive but I am still getting the same eht0 error. Rebooted the server, router and modem for a clean start and I'm still getting the device not found. Is there anything else to try before assuming my onboard ethernet is fried?
March 24, 201313 yr I see from your log that you have installed rc12. There was an issue with some NICs on that one and it was quickly replaced with rc12a. Maybe just bad timing on your part to grab a download with broken networking for your hardware. Try another download and see if that helps.
March 24, 201313 yr Author So, first I attempted to go back to the backup of my flash that happened right after the power outage. When I booted from that I was back to the unresponsive command line so it must have been running some recovery again... Then I went back to the copy of my flash that was known working from this morning. I booted successfully there so I powered down and replaced bzroot and bzimage with the 12a files. Booted back up and although unraid is responding to keyboard commands the eth0 adapter is still not found. the port itself has a solid red act/link light and an orange speed light. WAIT hold the presses... I just booted again (after having removed and replaced my USB to copy over a new syslog) NOW when I run ifconfig I have an IP. I can also access the server via the web GUI. And eht0 finds the adapter.... It seems like I'm back up and running. I also noticed from the web GUI that the last parity check was completed today so that unresponsive command line experience must have been a parity check running overnight. I'm surprised that's never been discussed here on the forums but I couldn't find any discussion of mdrecoveryd in relation to a parity check after an unexpected shutdown. So, fingers crossed... maybe everything is ok...
March 24, 201313 yr Author Also, while I'm getting back to square one here.... The flash drive I'm running right now is a fresh install (except for my config directory) Is it possible to copy over my packages directory or would I be better off just re-installing those manually? All I had installed were the basic programs covered by the configuration tutorial... unmenu, screen, preclear, etc... Also, the only other missing directory is the preclear_reports... Should I copy that over? Is this just for my records and for troubleshooting or does the system reference those preclear_reports for anything?
March 24, 201313 yr In addition to a parity check being triggered by an "unclean" shutdown, it's also possible that reiser journals are being replayed.
March 24, 201313 yr Also, the only other missing directory is the preclear_reports... Should I copy that over? Is this just for my records and for troubleshooting or does the system reference those preclear_reports for anything? They are for your own reference only. I would copy them over, even though the system does not use them, you might later see how a drive's syslog report has changed..
Archived
This topic is now archived and is closed to further replies.