This night Vera crashed at 3:00. I assume it’s because the time was set +1 hour.
(green light was flashing)
Did this happen to anyone else?
Is this device really that buggy?
I just checked and all looks fine on my system the DST was correctly set. (Vera 3 UI5 , version .622)
Cor
Yes, that happened to two of my three VeraLites.
Also often happens at the end of a calendar month… who knows why?
Generally tons of free and available memory.
Anyway, the occasional restart is good for the system :
Thanks for the feedback.
Unfortunately my Vera didn’t recover. I don’t know why.
[quote=“akbooer, post:3, topic:180452”]Also often happens at the end of a calendar month… who knows why?
Generally tons of free and available memory.[/quote]
Happened again at midnight last night, at the turning of the month… All three Veras.
About 15Mb free memory and between 30 and 40 Mb available memory on all machines.
[quote=“akbooer, post:5, topic:180452”][quote=“akbooer, post:3, topic:180452”]Also often happens at the end of a calendar month… who knows why?
Generally tons of free and available memory.[/quote]
Happened again at midnight last night, at the turning of the month… All three Veras.
About 15Mb free memory and between 30 and 40 Mb available memory on all machines.[/quote]
Do you have any scenes, PLEG actions or plugin processes that would be triggered when the month changes? Crashes are also caused by log-jams in Vera’s run-action queue.
I don’t use PLEG, and nothing, to my knowledge, is in any way month-referenced. I have some end-of-day actions, very lightweight, mostly just resetting some accumulators. I confess I haven’t actually looked at the logs (I tend not to these days, just interested in high-level events and variable changes logged by EventWatcher or DataYours.)
Do any of your myriad of ultra-smart plugins create new files/folders or call-up some external data source based on date?
When EventWatcher is enabled to write to a log file (which only ONE of my three is) then it rotates weekly on the tickover of the Unix epoch week (which just happens to be on a Thursday.) However, all three machines do it and, looking back in the event logs, this happens monotonously reliably at the start of every month, be it a Thursday or not.
I used to have a separate development machine, but now I’ve grown my network, I don’t, so it’s hard to start ripping out applications (and anyway each iteration would take a month.) Time to get a new Vera? Oh no, that’s the topic of another thread 
Not a huge issue, but somewhat galling.
I put a ticket in, just to be able to retrieve the logs which had rolled over before I got to them, and got this reply:
I checked the logs from your unit but as far as I can see the restart is normal. It's a routine meant to check the good health of the system.
Can’t see that this is well-known or documented anywhere?
Is this like with Windows XP which you should reinstall every now and then to maintain the good health of the system. :-\
A further response from MCV, when I asked whether this was a documented feature:
No, unfortunately this is not yet documented, but I will check with our development, so we can document it also.
Good find sir