June 28, 20233 yr Ever since my update I have to reboot my server every few days. Docker containers wont update even thought the UI is telling me there is an update. If a container is not running and I try to start it I just get the "Server Error" window that pops up. A reboot solves it but only last for a couple of days. Anyone have this happening? Any suggestions?
June 28, 20233 yr Yes I have been having what appear to be the same problems... but I can't even get a couple days, maybe several hours. Every morning though my sever is dead in the water again, usually noticeable by non-responsive Docker containers (espeically `HomeAssistant` which automates many things), but also when my wife can't access media because Shares are unreachable. I run headless so I don't have a monitor connected and have had to resort to hard-resets to get it back up and running. The other night I actually had to restart it 5-6 times, and though the UI was responsive most of those times, each time something different wonky was happening with Docker -- sometimes Docker Service wouldn't start, sometimes it would look like it started but ALL Containers said `Not Available', one time I tried to update a container and it looked like it worked but then Docker Service died and failed to re-start. It's been extremely frustrating & annoying . . . I do have some diagnostics (attaching), but I'm reverting to `6.11.5` to see if stability will return! unraid-diagnostics-20230627-2238.zip Edited June 28, 20233 yr by CajunCoding
June 28, 20233 yr Welp the downgrade did not go well now I'm stuck with a slew of `DockerClient.php` errors on my Docker Page... And if I try to disable and then re-enable the Docker service I'm greeted with a "Docker Service failed to start" error requiring a reboot . . . Edited June 28, 20233 yr by CajunCoding
June 28, 20233 yr Community Expert 4 hours ago, CajunCoding said: I do have some diagnostics (attaching), Jun 27 22:06:47 Unraid kernel: BTRFS: error (device sdj1: state A) in __btrfs_free_extent:3067: errno=-2 No such entry Jun 27 22:06:47 Unraid kernel: BTRFS info (device sdj1: state EA): forced readonly Cache filesystem is going read-only due to corruption, recommend backing up and re-formatting.
June 28, 20233 yr Man that sounds rough. Sorry to hear. Everything has been so good for me until 6.12. I to am headless but have a IPMI option if you will. Haven’t had to resort to that though. The web UI still works flawlessly when I have my issues. Beardifully sent Tapatalk style.
June 28, 20233 yr @JorgeB yes, based on Forum searches I reached a similar conclusion.... I don't know or why it suddenly got corrupted... and was just reading a thread you help support on, so Thank You very for responding and supporting us all on the Forum! 🙏 Ugggg, this was the point to having a Cache Pool . . . for redundancy to self-heal? I'm currently using a Cache Pool with 2 X Samsung 1TB SSD drives with BTRFS. I have some extra SATA cables that I'm thinking about swapping -- just in case but I find it odd that a cable would suddenly fail after nearly 2 years and right after an update to v6.12.1? And I use a dozen or so Docker applications . . . I'm currently moving all data to my array per documentation (backing-up-the-pool-to-the-array) . . . . My questions now are: I read that there are (maybe historically) numerous issues with curruption using BTRFS & Docker, etc. Since I have to re-build my cache, are there any other recommendations that I should do? Is there any other valid file-system for a Cache Pool? Edited June 28, 20233 yr by CajunCoding
June 28, 20233 yr Community Expert In my experience btrfs is pretty reliable when used with reliable hardware, but since v6.12 now has zfs you can try that one instead, if you then also get problems with zfs in the future it would suggest a hardware problem.
June 29, 20233 yr I'm having the same issues. I looked to see if my cache were going to read only, but so far I'm not having that issue. I'm attaching my diagnostics after a reboot, and trying to update a docker. Thank you! ETA: I can write to cache, so I don't know what's going on. tower-diagnostics-20230628-2056.zip Edited June 29, 20233 yr by camprman
June 30, 20233 yr Oh 6.12 has zfs option for Cache Pool? Unfortunately I've already rebuilt it... and am now trying to get my Dockers workign again... After running "Install Previous Apps"... they did NOT magically come back to life... currently Swag cannot load Certs, ports are all wrong, HomeAssistant isn't even loading data from `appdata` because it thinks it's a brand new install..... what a cluster! Now I'm trying to use the CA Backup/Restore (which I was faithfully running weekly) to see if that helps... it says that "No adjustments to the Docker templates should be necessary", but that's if you re-add Docker apps after you restore App Data. I had moved my share to my array and then re-set it back to Cache:Prefer so the mover would then restore `appdata`, but that did not work (soooooo Frustrating).... Edited June 30, 20233 yr by CajunCoding
June 30, 20233 yr Ok, Good News, I'm Finally back up and running! I deleted and re-created the Docker Image (Just to be safe), and I replace my SATA cables going to my two cache drives (also just-to-be-safe; and I may have gotten another CRC error but hard to tell if it was due to all the jockying around to get the Cache backed up (just-in-case), etc. But, none of that got me up and running . . . because I was still seeing Docker behaviour I described above where the Docker apps appeared to be ignoring the template settings -- primarily the Network settings. The Dockers were starting up (but numerous errors in logs), and most obviously many of the Networks said `None`. So re-created the custom network (I used a `proxynet` like many others based on SpaceInvaderOne's tutorials), and then the Docker Template reflected that setting, but alas was still showing `None` on the Docker list, and the wrong network IPs, etc. At this point no amount of stopping, re-started, deleting, re-installing was doing any good! Then I stumbled upon this Reddit post (Reddit can occasionally be super helpful) which said they toggled their Network to something else (other than the custom network) and then back again! Boom! FINALLY the Docker app loaded the correct settings, correct IP mappings, etc! `Swag` started right up without any errors in logs now -- perfectly happy to use `appsettings` and find my certs... `HomeAssistant` started up now without any issues also happy to use `appsettings` connect to MariaDB and start up just as expected... Of course `MariaDB` had the same treatment and responded similarly and is running fine now . . . Nextcloud had some hiccups, but logs showed it was installing an update, and then a few support searches later I was able to fix the 'stuck in maintenance mode' issue and now it's working great too! SOOO, all this to say that the "Install Previous Apps" does not work as well as we all wish it did -- especially if you have a custom network, etc. NOW, all this was triggered right after my update to 6.12 -- and others are experiencing similar issues. So I think I'll sit tight on 6.11.5 and not rock the boat for a (probably long) while!
June 30, 20233 yr I hope @Beardiful gets something helpful out of all this, and/or others keep chiming in and posting diagnostics so we can see if there may be an underlying issue....?
June 30, 20233 yr Reading your post makes me think I will try toggling my network settings. Thanks. I will report back. Beardifully sent Tapatalk style.
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.