@kartcon
Great discovery …
Did you happen to keep one of the files … to show to Vera ? Or publish here ?
I have a restart that happens when ever there is activity after some time of idle.
I have thought that it has to do with send an event to the Vera servers, but the connection has timed out, and it reloads to reinitialize the connection.
I plan to find a way to shutdown the event server and see if that help … I do not look at the Vera Events anyway.
Richard,
No, I didn’t think to save any of the files before removing them, but they were basically a text version of events typically seen in the Vera logs. Each was named numerically - presumably with some version of epoch time. The good news is in the past 24 hours I have had only one unexplained restart. That’s a frequency I can live with. I will consider this problem fixed if we can go a couple more days without another meltdown. Fingers crossed.
Please let me know I you can isolate the cause of the restarts on your system. I would love to get to the point of only having a restart when I schedule one. I suspect there is some common factor, but finding the cause may not be easy especially if detailed accounts can’t be provided.
Happy for you that you “discovered” what caused the issue and fix it. I checked the folder you mentioned to be empty on my VeraPlus. While following this thread I also checked restarts on the VeraPlus. Last restart was December 12th so pretty stable. In general I would say the VeraPlus restarts once every 6 days which is acceptable to me. Today it will restart a lot as I’ am going to get Alexa and play around with that. First thing to do rename all my devices to Engish (I’ am Dutch)… many restarts…
My VeraPlus restarts exactly every six minutes. I will open a ticket with Vera to see if they can figure out the problem.
I would also look at any device that could cycle in a 6 minute period (motion sensors, timers, schedules, scenes or any other device that ‘sleeps’ for a duration). If this is a new issue, I would go back to a prior backup and see if that solves the problem. If its been ongoing for 24 hours or more, start thinking about your most recent changes (new plug-ins, new scenes, devices or code).
Tech support is good, and in some cases the only way to fix a problem, but I also believe that as end users we are more intimately aware of the state of our system and need to take some responsibility for its condition. I’m not saying “You broke it, you fix it!” but I am saying in an open platform such as Vera, there are lots of ways to break things and we need to understand what progression of events brought us to this point. Sometimes simple logical thinking can shed light on the cause.
Good luck getting your system stable and please follow up once the issue is corrected. I’m sure there are many users who could benefit from your experience.
I think you are seeing the 6 minute checkpoint user data save. It looks like an error, but isn’t, e.g. and will look something like this:
02 01/02/18 16:51:11.199 UserData::TempLogFileSystemFailure start 0 <0x77489520>
02 01/02/18 16:51:11.219 UserData::TempLogFileSystemFailure 4606 res:1
-rw-r–r-- 1 root root 33 Dec 28 14:28 /etc/cmh/HW_Key
-rw-r–r-- 1 root root 32 Jul 3 2017 /etc/cmh/HW_Key2
-rw-r–r-- 1 root root 9 Dec 28 14:28 /etc/cmh/PK_AccessPoint
-rw-r–r-- 1 root root 7 Dec 28 14:28 /etc/cmh/PK_Account
…
I’d suggest installing the System Monitor plug-in, which will show you actual reboots and LUUP restarts.
My Vera 3 restarts several times a day, last time about 4 hours ago. But as you can see, Vera has not restarted in a while, not since I rebooted.
Don very kindly posted a screenshot of the output of System Monitor, which I should have done.
#$%@ I thought UI7 was finally stable, and bought a VeraSecure on Black Friday. I am not in the process of copying over configurations from Vera3/UI5, but tempted to just pull the plug out of the VeraSecure and let it collect dust for a while.
[quote=“RichardTSchaefer, post:21, topic:197952”]I have thought that it has to do with send an event to the Vera servers, but the connection has timed out, and it reloads to reinitialize the connection.
I plan to find a way to shutdown the event server and see if that help … I do not look at the Vera Events anyway.[/quote]
I assume you have all your devices set to disarmed? I know in the past we thought that prevented a device triggering from sending the event to the Vera servers…
FWIW To find out why my Vera was restarting, I went into apps/develop apps and added in the following to my startup Lua:
local filename = '/etc/cmh/restarts.log'
local outf = io.open(filename, 'a')
outf:write(os.date('%Y-%m-%d %H:%M:%S '..os.time()..' '))
outf:write('\n')
outf:close()
I then ssh’d onto the Vera, and checked “cat /etc/cmh/restarts.log” (note this file will continue to grow so will need to be monitored) on a regular basis. When I identified a restart, I then checked /tmp/log/cmh/LuaUPnP.log (note this file gets wiped when Vera reboots, as opposed to luup restarting) to find out what was happening at the time of the restart. In my case the restarts were happening directly after datamine tried to read data from a device. I removed that device from the list of things to log, and I found it still kept trying to read that same device but from a different device, and was still causing a restart. Very odd - basically something was corrupt. I decided I needed a reliable Vera more than I wanted historical data so removed datamine.
I didn’t do such deep investigation as you, but I have very similar experience - frequent luup restarts are (most probably) connected somehow to dataMine2 plugin, especially where there are more activity on the monitored devices.
I see also difference between Plus and Edge - on Plus, when I do some change on the device (i.e. name, or something in parameters) usually shortly after I have blue strip on top communicating about Luup restart. This doesn’t happen on Edge, at least not that often. I’m not sure if this is just difference in the reporting (like Edge is just not showing such information), or there is something different in real behaviour of both.
I’ve installed system monitor plugin to track restarts, but too recently to have any reliable conclusions.
[quote=“ninkasi, post:30, topic:197952”]FWIW To find out why my Vera was restarting, I went into apps/develop apps and added in the following to my startup Lua:
local filename = '/etc/cmh/restarts.log'
local outf = io.open(filename, 'a')
outf:write(os.date('%Y-%m-%d %H:%M:%S '..os.time()..' '))
outf:write('\n')
outf:close()
I then ssh’d onto the Vera, and checked “cat /etc/cmh/restarts.log” (note this file will continue to grow so will need to be monitored) on a regular basis. When I identified a restart, I then checked /tmp/log/cmh/LuaUPnP.log (note this file gets wiped when Vera reboots, as opposed to luup restarting) to find out what was happening at the time of the restart. In my case the restarts were happening directly after datamine tried to read data from a device. I removed that device from the list of things to log, and I found it still kept trying to read that same device but from a different device, and was still causing a restart. Very odd - basically something was corrupt. I decided I needed a reliable Vera more than I wanted historical data so removed datamine.[/quote]
Here is another way to do it - you don’t even have to ssh to pull up the “crash log”…
http://forum.micasaverde.com/index.php/topic,31292.msg226026.html#msg226026
Nice one - better than my hack… 
For me it is unfortunately more difficult, as I have one unit at remote location, where I’m present once per month, more or less, and don’t have much time to sit and track Vera behaviour.
Any Ideas how to access logs in such case?
By the way, I use DataMine to track memory reported by System Monitor plugin. And I’ve found something which seems to be interesting when I compared Edge and Plus data - amount of free or available memory on Edge have very high jumps up and down compared to Plus, for no apparent reason.
When you see the log it look as on the screenshot below
By comparison, memory on Vera Plus is much more smooth, regardless that it has about three times more devices installed. There are also jumps and downs, but within 5k of memory, while in Edge’s case it is about 20k of memory. From the other hand Plus memory goes up and down more frequently (there is jump “up” in free memory c.a. every three hours, while on Edge it happens every 6-8hours).
My Vera 3 reloaded LUUP around 8:30 this morning. Here is my memory graph:
In my case these jumps are not related to luup reloads. There was no luup reload during all period shown on the Edge’s graph which I’ve posted above.
There is a housekeeping process which runs every few hours to free up memory… This is OpenWRT level I think - it was there in previous firmwares (UI5, etc)…
It seems that something is wrong with this process on Edge.
And it seems that it has something to do with Datamine2.
Yesterday my Edge stopped responding and needed manual reboot. After this process all datamine channels were cleared and Datamine is ignoring USB memory attached to the controller (the memory shown on the Datamine Panel is Vera memory, not the USB stick).
And according to System Info plugin, there is much more memory free after that.
Seems that Datamine was buffering data on internal memory instead of saving it to USB stick, and this caused so high jumps in amount of memory available, compared to readings from Plus.
I uninstalled DataMine after reading several post here pointing to that. For me, that didn’t change anything.
My VeraPlus still restarts 10-12 times a day (Luup engine restart). I now log it to Syslog to be able to track it.
The restarts every second hour or so happens regardless of sensors being tripped or any other device activity. It happens when nobody is home too. I have several plugins that are active doing stuff, like Heliotrope, Weather Underground, Squeezebox and DEMII. Will continue to investigate and see if I can find a pattern.
[quote=“Thorden, post:39, topic:197952”]I uninstalled DataMine after reading several post here pointing to that. For me, that didn’t change anything.
My VeraPlus still restarts 10-12 times a day (Luup engine restart). I now log it to Syslog to be able to track it.
The restarts every second hour or so happens regardless of sensors being tripped or any other device activity. It happens when nobody is home too. I have several plugins that are active doing stuff, like Heliotrope, Weather Underground, Squeezebox and DEMII. Will continue to investigate and see if I can find a pattern.[/quote]
Try the code I linked to above that automatically grabs the last ~30 seconds of the luup log before a crash to help ID the problem. While I am new to UI7, I troubleshot my Vera3/UI5 extensively and found the problem it had with restarts using that code. You can do it straight from the log, but it is easier if you use something to grab it for you. Also see ElVira.xls spreadsheet process that you can use to dump the log around a crash (Luup engine restart) direct to Excel… [url=http://forum.micasaverde.com/index.php?topic=30265.0]http://forum.micasaverde.com/index.php?topic=30265.0[/url]