tzotz Posted February 9, 2017 Share Posted February 9, 2017 I have been using unRaid for a long time (since 2006) and am perfectly happy with it. I stayed with 4.7 and did not upgrade because it answered my needs. Today I couldn't connect to the web interface - received "this site can't be reached". Hoping for a graceful shut down by my UPS (APC SUA1000I) I turned power off and when nothing happened after quite some time I pressed the power button on the until until the system powered itself off. After that I couldn't get the system to function - I get as far as the Linux "Login:" prompt but that's it. I have almost zero Linux knowhow and can't get past this point to even begin to know what to do. I did copy the syslog to the boot device but when I looked at it on my Windows machine there were 5 logs dated today. I attach the first one (by time) thinking it may provide more info about what caused the problem couldn't attach more attachment size limit is exceeded. Please help anyone, thanks. syslog-20170209-100926.txt Quote Link to comment
Helmonder Posted February 9, 2017 Share Posted February 9, 2017 It looks like your array was running and even performing a parity check.. Where your shares reachable ? You have drive issues according to your SMART monitoring: Feb 9 10:09:00 Tower status[4210]: SMART overall-health self-assessment test result: FAILED! Feb 9 10:09:00 Tower status[4210]: Drive failure expected in less than 24 hours. SAVE ALL DATA. Feb 9 10:09:00 Tower status[4210]: Failed Attributes: Feb 9 10:09:00 Tower status[4210]: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE Feb 9 10:09:00 Tower status[4210]: 5 Reallocated_Sector_Ct 0x0033 105 105 140 Pre-fail Always FAILING_NOW 756 Quote Link to comment
tzotz Posted February 9, 2017 Author Share Posted February 9, 2017 Thank you. I can't connect to the web interface from my Windows Desktop nor get a reply when I try to ping the unRaid box. I can "talk" to unRaid through Linux terminal but I regretfully I'm almost completely "linux challenged". What should/could I do next? Quote Link to comment
Helmonder Posted February 9, 2017 Share Posted February 9, 2017 If you do not get a ping but you can "talk" to the system, how do you "talk" to the system, thru a keyboard and monitor connected to the unraid system ? Do you still have your shares ? Or they also unreachable ? Quote Link to comment
Helmonder Posted February 9, 2017 Share Posted February 9, 2017 Log on to the system (userid ROOT and the password you setup for that) Now give the following command: ifconfig You will get something back looking like the following: br0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.1.5 netmask 255.255.255.0 broadcast 0.0.0.0 ether 00:25:90:78:d5:5b txqueuelen 1000 (Ethernet) RX packets 994214 bytes 1364611213 (1.2 GiB) RX errors 0 dropped 635 overruns 0 frame 0 TX packets 637060 bytes 11386082379 (10.6 GiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0 You will possible see a few blocks of text, you need to find the one starting with eth0 or eth1 (where it says br0 in the above example), if you have found those blocks check what is listed after "inet" (192.168.1.5 in the above example). That should be the ip-address you recognize as the unraid one.. Also: When you are logged on to the unraid system, do a PING to your pc, see if that works. The syntax is the same as with windows/dos: ping <ip address> It kind of looks like your system has no network connection anymore, doublecheck all cables and connections and reboot... Quote Link to comment
tzotz Posted February 9, 2017 Author Share Posted February 9, 2017 Hi, thanks again. I did what you suggested and indeed it looks like there's no network connection. I'll investigate further and report. thanks Quote Link to comment
tzotz Posted February 9, 2017 Author Share Posted February 9, 2017 hallelujah, Thank you Helmonder, It was, after all a network (apparently hardware) problem: - resetting the network cable connectors didn't work - replacing the cable from the box to the wall socket didn't work - using an alternate network card (luckily there were 2 NICs on this MB) and lo and behold: everything works like new. I will run a parity check in light of the drive SMART warning. Thanks a million for your help - made my day. Best regards, Tzotz Quote Link to comment
Helmonder Posted February 9, 2017 Share Posted February 9, 2017 hallelujah, Thank you Helmonder, It was, after all a network (apparently hardware) problem: - resetting the network cable connectors didn't work - replacing the cable from the box to the wall socket didn't work - using an alternate network card (luckily there were 2 NICs on this MB) and lo and behold: everything works like new. I will run a parity check in light of the drive SMART warning. Thanks a million for your help - made my day. Best regards, Tzotz The parity check will not help.. One of your drives needs replacing.. When Smart tells you a disk is FAILING_NOW it means it is failing now... Get a replacement disk and preclear it asap so you can put it in.. Quote Link to comment
tzotz Posted February 9, 2017 Author Share Posted February 9, 2017 will do. How does one preclear? replace the drive and then what? Thanks again Quote Link to comment
Helmonder Posted February 9, 2017 Share Posted February 9, 2017 Oeff.. it has been a while since you have used unraid.. Preclear is an extensive test of the drive, to make sure that a drive you just bought is actually good before you start using it.. It is a commandline based tool.. In the V6 unraid versions there are plugins that make it easy to do this thru the web-gui, but for 4.7 you need to do it via command-line.. Search the forum for the software.. You need PRECLEAR and SCREEN. Then follow the following instructions: ----------------------------------------------------------------------------------------------------------------------------------------- Preclearing Multiple pre-clearing cycles are good, 3 is a good number. They will however take a lot of time. Note that if you start preclear by telnetting into the server from your pc the preclearing is dependent on your pc continuing to work (a reboot of your windows box will stop the preclear, this can be prevented by using "SCREEN" (it will make sure the preclear will continue even with your windows box shut down) The following commands will have you covered on both (3 cycles with screen and sending reports on progress by email): cd /boot screen ./preclear_disk.sh -w 65536 -r 65536 -b 200 -c 3 /dev/sd? (replace the ? with the correct letter) You can now type CTRL-A and d to "detach screen", you can now safely exit from your terminal window, the preclear will continue. To get back your preclear screen, open a telnet session and type SCREEN -r <id> ----------------------------------------------------------------------------------------------------------------------------------------- Quote Link to comment
trurl Posted February 9, 2017 Share Posted February 9, 2017 After you get your system stable PLEASE consider upgrading, if your hardware will support V6. See the Upgrading sticky linked in my sig. Even if you only want the NAS functionality, V6 is a much better NAS than previous versions. It will tell you when you start to have disk problems so you can take care of them before it's too late. And you will get much better support on the forum. Most of us haven't seen your version for very many years. Quote Link to comment
tzotz Posted February 9, 2017 Author Share Posted February 9, 2017 about preclearing: When I insert the disk to be precleared in the array, does it get the same device code as the one that it replaces e.g "sdg"? As I understand, preclearing takes place with the array active. How does one guard against preclearing the wrong disk? disconnect all the disks except the one? thanks Quote Link to comment
Helmonder Posted February 9, 2017 Share Posted February 9, 2017 You need to give the drive name in the preclear command.. ./preclear_disk.sh -w 65536 -r 65536 -b 200 -c 3 /dev/sd? (replace the ? with the correct letter) So you control that for yourself. There is some logic in the preclear system to avoid you preclearing an existing drive but you should not trust on that. Just check what is the name of the drive you want to preclear. Quote Link to comment
RobJ Posted February 9, 2017 Share Posted February 9, 2017 To find the current drive symbol (sdX or hdX) for the drive you want to Preclear, use: preclear.sh -l (that's an el, not a one) Quote Link to comment
JonathanM Posted February 9, 2017 Share Posted February 9, 2017 I have to respectfully disagree with the insistence on preclearing using the at risk unraid server in this specific situation. I think the OP's time would be better spent letting the drive rebuild and watching for errors, then doing a non-correcting parity check to verify the rebuild went ok. Checking the new drive in another box using the manufacturers tools while leaving the unraid server off would be better IMHO. Too much room for error trying to install the scripts and learning to preclear a drive while the array is in a degraded state. Quote Link to comment
tzotz Posted March 3, 2017 Author Share Posted March 3, 2017 Hi again, Progress made but some questions (trivial I hope) still remain: Operations under 4.7: Bought a new (bigger) disk Precleared it (twice) on an offline machine (to be on the safe side), Replaced the failing disk, Powered up unRaid and rebuilt the new disk, 0 errors. Restarted the array - everything AOK Stopped the array then powered down unRaid V6.3.2 Migration: Prepared the flash for V6.3.2 migration as per instructions in "Upgrading_to_UnRAID_v6" Powered up unRaid with the new flash using the default boot option (GUI OS) Activated the web interface (http://tower) Assigned the disks as per the previous disk map. A red notification says "Parity Disk Will Be Rewritten". Please see attached 6.3.2_main.jpg Questions: (1) Is there a way to verify the validity of the array without rewriting any disk? (2) If I select "Parity is already valid" then start the array without parity check - should it not be valid since I stopped 4.7 before starting 6.3.2? Moving on to Disk Settings. Please see attached 6.3.2 VS 4.7 Disk Setting.jpg Questions: (1) Default file system is set as XFS , my 4.7 system was ReiserFS - I assume this is the correct value. (2) A whole bunch of parameters are set to their default values, some are not identical to the previous 4.7 values and some are new, should I leave them all at default? I'm not going to start the array and continue setting it up until I'm confident that everything is OK up to now and that my data won't be compromised. Thanks. Quote Link to comment
garycase Posted March 3, 2017 Share Posted March 3, 2017 First, be CERTAIN that you have your disks assigned correctly -- i.e. you have not mistakenly assigned a data disk as parity. This would result in the loss of all data on that disk if you assigned the wrong disk. As long as the disk assignments are correct, AND you are CERTAIN that you made NO changes to the array (i.e. no writes at all) between shutting down the v4.7 config and setting up the flash drive with 6.3, then yes, you can check the "parity is already valid" box and it should be okay. I'd leave the default file system set to XFS. This has nothing to do with your current disks -- it just means if you add a new disk to the array it will be formatted with this file system. There's no problem mixing RFS and XFS disks in the array. Quote Link to comment
JonathanM Posted March 3, 2017 Share Posted March 3, 2017 A quick caution. If you follow all the good advice so far and start the array, there is a minor possibility that one or more of your correctly assigned RFS drives show up as unformatted. If that is the case, don't panic. Stop the array, and change the file system type ONLY on the effected disks to RFS. Leave the default as XFS, it won't change the corner case that may cause your drives to appear unformatted. Once you change the individual drive to RFS, it will no longer attempt to make a guess about what file system type, and should come up as RFS without any issues. The issue is that sometimes the auto file system setting fails at correctly detecting the existing format, so setting it manually to the correct type is required. Quote Link to comment
garycase Posted March 3, 2017 Share Posted March 3, 2017 One other thing to remember -- once you've got the array loaded, you'll probably need to run the "New Permissions" utility (on the Tools tab) one time. Quote Link to comment
tzotz Posted March 3, 2017 Author Share Posted March 3, 2017 Thanks guys, I appreciate your help. garycase I won't forget about the "New Permissions" utility jonathanm, thanks for the heads up about the FS, BTW having read that I'm panicked already, so thanks for your reassuring words. Questions: Does the "unformatted" notification appear (if at all) after I power up unRaid or after I click on Start Where's the individual disk setting? In Settings under "Disk Settings" I only see the global settings or do I? Thanks again, Quote Link to comment
garycase Posted March 3, 2017 Share Posted March 3, 2017 25 minutes ago, tzotz said: Does the "unformatted" notification appear (if at all) after I power up unRaid or after I click on Start After you click on Start. 25 minutes ago, tzotz said: Where's the individual disk setting? In Settings under "Disk Settings" I only see the global settings or do I? On the Main display tab, if you click on a disk (e.g. "Disk 1") it will bring up a page of info on the disk -- the file system type is one of elements on that page. Quote Link to comment
garycase Posted March 3, 2017 Share Posted March 3, 2017 Basically, you're ready to go ... just click on the "Parity is already valid" box and Start the array ... unless it shows an unformatted disk all should be well. If by chance it shows an unformatted disk, do NOT check the box that indicates it will format the disk; just change the file system type as jonathanm noted above. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.