062bel313 Posted May 29 Share Posted May 29 Hello all: my unraid web gui is un responsive. I can log in one time which takes me to Main. After that I cannot do anything, cannot go to different menu like Docker, Dashboard. I can go to Settings and tools to get the diagnostics. Can you point me in the right direction? I have attached the diagnostics from the unraid. Thank you. unraid-diagnostics-20240529-0437.zip Quote Link to comment
itimpi Posted May 29 Share Posted May 29 The syslog shows that you have corruption on disk2: May 28 08:42:26 unraid kernel: XFS (md2p1): Metadata corruption detected at xfs_dinode_verify+0xa0/0x732 [xfs], inode 0x30e48b5eb dinode May 28 08:42:26 unraid kernel: XFS (md2p1): Unmount and run xfs_repair If you can you should run a check filesystem on that drive. It may also be a by-product of an underlying hardware issue such as RAM playing up. 1 Quote Link to comment
062bel313 Posted May 29 Author Share Posted May 29 I am trying to stop the array in maintenance mode from GUI, but the proceed doesn't do anything. Is there option to stop it via SSH? Quote Link to comment
062bel313 Posted May 29 Author Share Posted May 29 I was able to go to maintenance mode by rebooting the system to safe mode. I ran >> xfs_repair -v /dev/md2p1. Looks like it cleared reflink flags on inode during Phase 4. Not sure what it meant but the diagnostic in syslog no longer shows any issue. I will keep the system on for few days and see if the issue revisits. I will update again in few days with the progress. Thank you @itimpi Quote Link to comment
itimpi Posted May 29 Share Posted May 29 When you rebooted in normal mode did the drive mount OK? Was there a lost+found folder created which is where the repair process puts files for which it does not know the name. 1 Quote Link to comment
062bel313 Posted May 29 Author Share Posted May 29 Yes the drive mounted OK. There is no lost+found folder in the root of boot. Is that the location I should check? Quote Link to comment
itimpi Posted May 29 Share Posted May 29 Just now, 062bel313 said: Is that the location I should check? No - it would be on the drive you repaired (and show up as a share). Not having it is a good sign as it means that there was probably no data loss at all. 1 Quote Link to comment
062bel313 Posted May 30 Author Share Posted May 30 I checked and disk 2 doesn't have any Lost+Found. Web UI is accessible so far. Some how Parity check ran, may be because there was some issue in reboot. Parity check going about 230MB/sec will take few more hours. Hopefully things will get to normal. 🙂 Quote Link to comment
062bel313 Posted May 31 Author Share Posted May 31 Now the xfs corruption shifted to disk 1 😢. Going to do same thing for disk 1. Quote Link to comment
JorgeB Posted June 1 Share Posted June 1 In case you missed it there's also fs corruption on disk2. 1 Quote Link to comment
062bel313 Posted June 1 Author Share Posted June 1 (edited) I noticed that, so I ran xfs_repair for both disk 2 before and still see that. I could not find lost+found in disk root directory. Am I using the right syntax: xfs_repair -V -L /dev/md2p1 xfs_repair -V -L /dev/md1p1 My server crashed this morning. When checking in the UDM (router), I was not able to see unRAID IP in the clients, but server was running; running in the sense that it had power but was not able to access it via SSH and WebUI. so had to force restart. Not sure what is causing this crash. Here is the diagnostics after the restart, if you can find something. unraid-diagnostics-20240601-1108.zip Edited June 1 by 062bel313 Quote Link to comment
JorgeB Posted June 2 Share Posted June 2 16 hours ago, 062bel313 said: xfs_repair -V -L /dev/md2p1 Should be first xfs_repair /dev/md2p1 Only use -L if it asks for it 16 hours ago, 062bel313 said: Not sure what is causing this crash. Here is the diagnostics after the restart, if you can find something. Syslog starts over after every boot, you can enable the syslog server and post that after a crash to see if it catches something. 1 Quote Link to comment
062bel313 Posted June 4 Author Share Posted June 4 (edited) Unfortunately, when I was planning to enable syslog server, I noticed that the server is not accessible from web UI as well as ssh. I did hard reboot and unraid is seen in the DHCP clients of my router but cannot access webui and ssh. Looks like the server is dead. 😭 What should be my next step, any guidance? Edited June 4 by 062bel313 wrong information Quote Link to comment
JorgeB Posted June 4 Share Posted June 4 2 hours ago, 062bel313 said: Looks like the server is dead. 😭 What should be my next step, any guidance? Do you still get the CLI or is it dead dead? 1 Quote Link to comment
062bel313 Posted June 4 Author Share Posted June 4 no cli. cannot ssh to the machine. it is headless, may be try to hook the monitor? Quote Link to comment
JorgeB Posted June 4 Share Posted June 4 5 minutes ago, 062bel313 said: may be try to hook the monitor? Yep. 1 Quote Link to comment
062bel313 Posted June 4 Author Share Posted June 4 in the monitor i see: unraid login: sed: couldn’t open temporary file /boot/config/sedwmnHkH: Read-only file system but this time i was able to login the ui, and i see red notification USB flash drive failure: time Alert[UNRAID] - USB drive is not read-write Cruzer_Fit(sda) does this mean my usb is failing? I will enable syslog server this time. Quote Link to comment
062bel313 Posted June 5 Author Share Posted June 5 Did a chkdsk of a boot drive on a windows machine fixed the read-only issue. I was also able to enable the syslog mirror on the usb drive. Hopefully I will be able to share the diagnostic now if server crashes. 1 Quote Link to comment
062bel313 Posted June 6 Author Share Posted June 6 I am back, but server hasn't crashed 🤩..... yet. So I am in an aggressive monitoring mode. I am checking the syslog to see if I encounter any lines with red color or mention of error. I see there is some lines with kernel tainted in CPU 3 with some proccess ID lsof. I have no idea what it means, a quick google and my speculation suggests that the kernel is not in pristine condition. Can you check with your experienced eyes if you see any warnings or errors, that I need to worry? The diagnostics are attached. unraid-diagnostics-20240605-2147.zip Quote Link to comment
JorgeB Posted June 6 Share Posted June 6 There's a call trace, start by running memtest, or re-test with just one stick of RAM, if the same try the other one, that will basically rule out a RAM issue. Quote Link to comment
062bel313 Posted June 7 Author Share Posted June 7 Did a memtest from unraid boot option with both RAM. passed the test for 6 passes. Quote Link to comment
JorgeB Posted June 8 Share Posted June 8 Since memtest is only definitive if it finds errors, On 6/6/2024 at 9:23 AM, JorgeB said: re-test with just one stick of RAM, if the same try the other one, that will basically rule out a RAM issue. Quote Link to comment
062bel313 Posted June 13 Author Share Posted June 13 As suggested, I did memtest. I have two RAM sticks. I did the test again for individual RAMs and all tests are passed. It took long. What I did was after 6 successfull passes I stopped and exchanged the RAM and did memtest again for 6 passes. Is that the process, or should I let it continue? I don't know how many passes it needs to confirm that RAMs are fine. After that I rebooted the machine to unRAID and I see there are trace calls (like before) as attached. Any suggestions? Quote Link to comment
itimpi Posted June 13 Share Posted June 13 The point is that running memtest is only definitive if it finds errors. What was suggested was that you run the server with less RAM sticks plugged in to see if you still get errors. not that you run memtest with less sticks plugged in. Quote Link to comment
062bel313 Posted June 13 Author Share Posted June 13 Oh got it, I will give it another try with less RAM and run unraid for a while. 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.