C4 Hub doesn't come up after soft reboot and takes 30 mins to boot after pulling power

Unfortunately, I'm having some new issues on .147. I'm getting occasional Error 500s when looking at my Rule Machine rules, and my zigbee network just went offline. While the free memory looks a lot better in .147, these issues are pretty significant.

I'm going to roll back to .145 so that my house doesn't fall apart during thanksgiving chaos tomorrow.

I'll grab the logs and investigate.

1 Like

Is this new issue corrected in .148? Or should I wait for a future release before upgrading again?

148 has a few useful fixes. I think during upgrade to 147 the previous version of firmware didn't completely stop or exit cleanly. So let's do this:

  • Run/download a backup and save it somewhere.
  • Update to 148, wait for it to come up.
  • Reboot the hub right after, just to be sure thing clear up.

I performed the steps you suggested, rebooted after the upgrade completed, and the hub still crashed about 2 hours later. I had to pull power on the hub to bring it online this morning, because the UI was unresponsive. Something changed in .147 to make my hub very unstable, and it persists in .148. You can see from the memory graph below that part of the hub was still alive at 3am, which is when it is scheduled to reboot but failed to do so.

I'm going to give it one more shot on this version and then I'll have to roll back to .145. Please let me know what I can do to gather more information for you so that this issue can be fixed.

Please stay on 145 for now since later versions are repeatably breaking. I'll grab the logs off the hub in the meantime.

Ok, rolling back now.

Should I open a separate ticket for this issue? I'd like it to get looked at before the code gets developed much further, as I'm sure it'll be harder to fix the longer we wait.

I can temporarily add your hub to the beta group, where the change that broke 147/148 for you is rolled back (and other things are fixed along with it). If you don't care for the beta, the change will be in the next general public release. Can't pin an exact date on it as a matter of policy, but we tend to be pretty good with the turnaround times.

As long as the fix has been identified and will be in the next release, I can just wait for it. I appreciate the options, though!

Do you know what build number the fix will be in? 149? Or will be be in an upcoming major release, instead of a hotfix release?

It will definitely be a hot fix for 2.2.4. We're not adding features until current iteration is stable.

1 Like

Would it be helpful to you if I entered the beta train and tested the fix? It seems easy to reproduce in my environment, as my hub hasn't been able to function for more than 3 hours on .147 or .148. I'm happy to do that, as long as I can roll back to a stable version and opt back out of the beta train once testing is complete.