Cannot update 2.2.5.126 to 2.2.5.127

@gopher.ny, @bobbyD, I’ve got a similar issue on my C-5, now trying to update from 2.2.5.126 to 2.2.5.127, that I reported two days ago on my C-7 updating 2.2.5.125 updating to 2.2.5.126. See: Cannot Update to 2.2.5.126, cannot add some z-wave devices to .125 / c7 hub

The C-5, my development hub, is lightly loaded, only one device. I test out releases before doing the update on my production C-7, which has a couple dozen devices.

Today, I tried to do the update from 2.2.5.126 to 2.2.5.127 on the C-5. I did a backup and locally downloaded 2.2.5.126, then started the update. It’s been stuck at the Extracting Update screen for over 30 minutes an hour.

Diagnostic Tools (port 8081) are unreachable, no response. Hub IP is unreachable.

Any suggestions? Time to pull the plug, revert to 2.2.5.126, soft reset, restore from 2.2.5.126 backup?

Did it change IP?

1 Like

+1 on that. Hub now reboots on version update, and router may assign it a IP while it's rebooting.

1 Like

No, but that was a good thought. There is a DHCP lease reservation for both of my hubs, C-5 at 10.0.1.200, C-7 at 10.0.1.209. Both have been at their same IP since they were unboxed.

One additional piece of info (besides still stuck at Extracting Update for 2 hours now): C-5 Hub light is still blue (as I would expect while Extracting Update).

I just checked my network with a scanner, just to make sure that, somehow, the DHCP system hadn’t screwed up and assigned something other than the lease reservation, and all devices on the network are accounted for. No new devices, still nothing at 10.0.1.200, the lease reservation. I believe it must have crashed during the update.

Any other suggestions other than pull plug, soft reset, revert to 2.2.5.126, restore backup?

Ok, in the absence of a reply, I pulled the plug, power down 30 seconds, applied power, it booted to green light at the assigned DHCP lease reservation (10.0.1.200), so it didn’t change IP, but instead just crashed. Shows as running 2.2.5.127 after booting.

@gopher.ny, do you want to look at it before I revert to 2.2.5.126, soft reset, restore backup, and update again?

I’m just a bit uneasy running on a DB after a crash. This hub has never before crashed since September 2019, when unboxed.

Not a single useful thing in the logs. It's as if power plug was pulled (which I'm sure didn't happen) and things just stopped. I can't say for sure what occurred, but I also don't think reverting is necessary.

In my case, my C5 does not show "update available " but my C7 does! I'm not using the C7 yet so I'll just wait until this is all sorted out.

I ran a "check for update" and it shows it has one, sorry for the confusion.

Ok. Thanks so very much for taking a look at it. Perhaps it was a cosmic particle from space. Both hubs are on a UPS (different UPS devices, different circuits), haven’t seen any power issues.

It’s only my development hub. It’s odd, though, that both of the last two updates have gone badly, similarly, but the previous time left port 8081 operational on the C-7.

Crossing my fingers, going for the update on my C-7 hub (backing up first, as always).

There's now a download latest backup button next to update button, a gentle reminder to back up :slight_smile:

3 Likes

Yes, I noticed that. Saves a step from first going to the Backup and Restore tab before doing the Check for Updates. Very, very nice! Things are getting better and better, and I appreciate your work. The one regression test I have not yet made on this recent flurry of updates is whether the date in the backup file name still shows as the GMT date rather than the local time date. I have to make a backup within 5 hours of local midnight to check that. It was still broken as of 2.2.5.126 (see screenshot). Note that last 2.2.5.126 backup says it was taken on Feb 3, which is in the future.

This is a hard bug to observe because the 2:15 am backup overwrites the single backup saved with that date on the hub.

FYI, the C-7 update from 2.2.5.126 to 2.2.5.127 went smoothly. Life is good.

This bug appears to have been fixed with 2.2.5.127:

Note that the backup was created an hour before midnight, when GMT would have been Feb. 3.

FYI, a “feature” of the “download latest backup button next to update button” is that this download doesn’t prepend the Hub name to the backup file’s name, as the “Create New Backup and Download Locally” “Download” button on the Settings > Backup and Restore page.

Not that it should be changed, because some people complained that prepending the hub name broke their automated backup scripts. I appreciate the prepended hub name on backups made on the Backup and Restore page, but just wanted to point out the difference.

Thanks for all you do.

FYI, today’s update from 2.2.5.128 to 2.2.5.129 went smoothly on my C-5 and C-7.

1 Like