5.14.x - Controller will randomly not start due to Unsatisfied dependency / Circular Reference
This was reported on my repo for the container image I build. This is happening for both the beta and GA versions of 5.14.x: both of the latest beta versions, 5.14.0.11 and 5.14.20.9 and the 5.14.26.1 GA version. There are reports of the app not starting. Oddly enough, it's a scenario where you can take the same Docker image, run it on two identical machines (same OS, same kernel, same Docker version) and get two different results. One will run it fine, the other won't. This is a brand new, clean installation with no previous controller configuration data. I will try to post the logs as a follow up but it keeps saying I am posting an "illegal link" in my post for some reason - probably something in the text of the post.
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
Same with me
Watchtower just upgraded my omada controller to 5.14.26.1 and it runs into the same issue. I have reverted back to 5.13.30.8 which is working fine.
- Copy Link
- Report Inappropriate Content
@mbentley Same with me Watchtower updated to 5.14 and no start, had to downgrade to 5.13. Now everything is working.
- Copy Link
- Report Inappropriate Content
Having the same issue as all above, updated to 5.14.26.1 and had to revert to 5.13, complains about some unsatisfied dep.
C'mon @TP-Lick ... pitter patter lets get at 'er!
- Copy Link
- Report Inappropriate Content
@mbentley I've the samme issue after upgrading to 5.14.
- Copy Link
- Report Inappropriate Content
@KOp same issue.. would be nice if TP-Link could resolve this bean issue.
- Copy Link
- Report Inappropriate Content
I saw the upgrade notification today and tried to upgrade and am also running into this issue with 5.14.26.1.
- Copy Link
- Report Inappropriate Content
I am just curious - is there anyone who has mentioned having the same issue who is running the controller directly on the operating system instead of using the container image? I would try to do a native install but it is really hard to know if it will not work on any given machine.
My biggest concern now is for the people who have been able to successfully run the 5.14 image because the problem appears to show up randomly. On my server, 5.14 runs fine but for all I know, it will suddenly stop working tomorrow and I will not be able to just simply roll back to 5.13 because it's already successfully ran 5.14.
I tested the container on 6 different machines, all using the exact same container image (mbentley/omada-controller:5.14@sha256:ce429a64b11bcc8f86fdc5aa1990a54a158de8ce83f92c316bb556f29f732afd) and it only failed to work on one of them.
I rebuilt the image without a cache (which is mbentley/omada-controller:5.14-rebuild-amd64@sha256:8cf6a498927076a49d56ec4ef202015e48951a313c822a99dd4a38dae09b8be9) and the exactly same machine had the error happen again. The other 5 were fine.
Just for kicks, I tried the 5.14.20.9 beta (which is mbentley/omada-controller:beta@sha256:2162c15ddb596fca42017613d968388631ee76a50f3e24203d56694349dab44c) and it works fine on all 6.
- Copy Link
- Report Inappropriate Content
same problem here after upgrading to 5.14
(also running your image, not directly on a machine. Rollback to 5.13 helped.)
- Copy Link
- Report Inappropriate Content
same problem here - had to get back to 5.13
- Copy Link
- Report Inappropriate Content
@mbentley running the 5.14.20.9 beta for a while now and works fine!
The 5.14.26.1 fails to start, same error as you described. Running on docker and amd64 version.
Rolled back to 5.14.20.9 beta and it works.
- Copy Link
- Report Inappropriate Content
Information
Helpful: 17
Views: 7075
Replies: 43
Voters 4
