January 4, 20188 yr Running unRAID 6.3.5 Originally, I was getting the (fantastically unhelpful) "Execution Failed" message when I would try to start any of my Docker containers after stopping them (at this point they seemed to be autostarting successfully on server startup). Following the prescribed methods, I blew away my Docker image file, re-created it, and then re-created all of my Docker containers. Everything seemed to be fine for a bit then I started to get the "Execution Failed" message again. Now, after restarting the server with no other changes made I cannot access the GUI any longer. When I try to access it, it just clocks indefinitely (no page appears, no login prompt). Docker Containers: Couch Potato (based on some of the stuff in the logs, a likely culprit), Radarr, Sonarr, PMS, MySQL, and SABnzbd. All from linuxserver.io, all set to auto-start except for MySQL. I can still Telnet into the box without any issue and all of my SMB shares appear to be fully accessible. None of the GUIs for any of my Docker applications come up, either, though they do not clock - they fail immediately. The Docker tomfoolery and the GUI nonsense aren't necessarily related but that sure seems unlikely. I know far too little about unRAID, Docker, or Linux to troubleshoot this further. Please help if you can! newbsauce-diagnostics-20180104-1314.zip Edited January 4, 20188 yr by MethodCall
January 4, 20188 yr Your cache filesystem is completely allocated resulting in ENOSPC errors, see here for how to proceed: https://lime-technology.com/forums/topic/62230-out-of-space-errors-on-cache-drive/?do=findComment&comment=610551
January 4, 20188 yr Author 10 minutes ago, johnnie.black said: Your cache filesystem is completely allocated resulting in ENOSPC errors I have *literally* no idea what that means...but I'll dig into that thread, as you suggest. Thank ya, sir.
January 4, 20188 yr Author I did have to nuke some downloads to make space for the process, but I have completed a balance -dusage=75 without an error. To anyone performing this in the future, as that -dusage percentage goes up... so does the time it takes to complete. That last one at 75% took *at least* 20 minutes. Restarting the server now to see if this did the trick. *crosses fingers*
January 4, 20188 yr Author Damn. No joy. GUI is still clocking infinitely. New Diagnostics ZIP attached. newbsauce-diagnostics-20180104-1534.zip
January 4, 20188 yr Forgot to add that you'll need to delete recreate docker image after the balance.
January 4, 20188 yr Author 1 minute ago, johnnie.black said: Forgot to add that you'll need to delete recreate docker image after the balance. Will deleting it via mc be sufficient? As far as re-creating it, I don't know how to do that outside of the GUI (which I still can't get back to) but if you can gimme the commands, I'll throw 'em at Putty. You think something with Docker is breaking the GUI? 1 minute ago, johnnie.black said: And make sure to reboot also I rebooted the box after the balancing, you want me to reboot after deleting the Docker image, or after deleting and re-creating the Docker image?
January 4, 20188 yr Try this: With mc navigate to and edit /boot/config/docker.cfg and change DOCKER_ENABLE to "no", reboot and if the GUI is working delete and recreate de docker image, if the GUI still doesn't work grab new diags by typing diagnostics on the console.
January 4, 20188 yr Author Edited & saved. Rebooted. Reopened docker.cfg after reboot to confirm that the edit took (it had). GUI is *still* clocking. Good grief, what have I done to my server? New diags attached... newbsauce-diagnostics-20180104-1602.zip
January 4, 20188 yr Author Interesting. (In a bad way.) Tried to shut down the server with powerdown (instead of the powerdown -r's I've been doing all along) to make absolutely sure it restarted. So I physically went over to the server (since I was gonna have to hit the power switch to turn it back on, obviously). It hadn't shut down. Reconnect with Putty...sure enough, still up. Tell it to powerdown again. Nothing. I mean, the CLI is telling me that it's going down, got the broadcast message like usual...but the front of the server's chassis is still a frickin' christmas tree. No actual shut down is taking place. So, I kill it by holding down the power switch. It goes down for real. Hit the switch again to power it up...wait a bit for it to post annnnnnndd.....the GUI is accessible again. I'm questioning whether or not ANY of my previous powerdown -r's ever actually rebooted anything or not, now. Lesson learned...get visual confirmation on shut downs. So.... delete and re-create my docker image, I'm assuming?
January 4, 20188 yr The powerdown command is obsolete. With unRAID6 use poweroff - to do a clean shutdown of the server reboot - to do a clean restart of the server
January 4, 20188 yr 5 minutes ago, MethodCall said: I'm questioning whether or not ANY of my previous powerdown -r's ever actually rebooted anything or not Not the ones from today, syslog started yesterday. 6 minutes ago, MethodCall said: So.... delete and re-create my docker image, I'm assuming? If all is working maybe not needed, but since it's easy and just in case I still recommend doing it.
January 4, 20188 yr Author Just now, bonienl said: The powerdown command is obsolete. With unRAID6 use poweroff - to do a clean shutdown of the server reboot - to do a clean restart of the server *facepalm* Hey, Lime....maybe an error message inside the CLI would more effectively communicate that....instead of the broadcast message that gives other-than-true feedback to the user. Just a thought. *sigh* Good info, Bonienl. That is *definitely* something I need to know. Appreciate it!
January 4, 20188 yr Author 4 minutes ago, johnnie.black said: If all is working maybe not needed, but since it's easy and just in case I still recommend doing it. Yeah, no more assumptions from me. DEFINITELY making sure. 100%.
January 4, 20188 yr Author It seems as though everything is back to normal operation. Man, I can't thank you enough, Johnnie. There is pretty much *zero* chance I would have figured out the problem without ya. If I could sling more upvotes your way, I would...but I have literally hit the daily limit. Top shelf, son. TOP SHELF.
January 5, 20188 yr Good, while on v6.3.5 it's good practice to schedule a weekly balance with -dusage=75 to keep the filesystem allocation under control, this issue it's supposedly fixed on the newer kernel used on v6.4, though still early to say if it's really fixed or not.
Archived
This topic is now archived and is closed to further replies.