Reloads: Why

Hopefully this is the right area. Again this month my Vera is restarting frequently. What should I be looking for in the logs to pinpoint the cause? Suggestions?

Cheers

C

Not to be too flippant, it does so because it is a Vera, and it is notorious for it.

We are all here hoping that this may be fixed in the next generation firmware.

However, for the time being, it depends how brave/capable you are in taking any of the mitigating approaches that are described elsewhere on the forum.

You may, for example, look at this

1 Like

Oh I agree it\s notorious but there’s no way it’s gone from no reboots in 3 weeks to rebooting every hour or less without something happening. That’s what I’m looking for.
It’s been exrooted and has loads of memory.

Rafale’s definition of instability was around the every couple of weeks mark, not this (again!)

Cheers

C

Quick analysis would be to go look for memory leak (try SSH in and try the free command) or storage running out of space. If you want to go deeper, you would need to go look for the luaUPnP logs which are in /tmp/log/cmh/ and see what error code you are getting.
It could be nearly anything…

Yeah, so memory looks fine.
Logs rotate fine every 12 hours up until 0512 local today.
Next rotate was at 0940 local
Virtually every rotated Lua log has rdate ($PID) Unable to sync time with exit_code=1

Then we have
LuaUPnP Terminated with Exit Code: (chose from 0, 137 or 245)

0 I guess is normal. 137 is kill -9 I believe. No idea what 245 is

Just happened again while typing this

Grrrr

C

Well score one for the iOS App

Sitting in the hotel bar watching it ping me alerts that it’sm restarting, I decide to try and turn the damned things off.

Alerts in the mobile App shows entry after entry of ‘Network Monitor is stuck’

RIgpapa did you kill this somehow? I know there’s nothing wrong with my network…

Can’t even find the process in top nor do I get any logs…

C

Didn’t you see this same problem with Network Monitor several weeks ago??

How do I figure into this?

You might look at @rafale77’s thread on taking vera of the grid. I believe that he modifies or disables NetworkMonitor.sh in his scripts.

Yes I got there in the end! I needed to re-enable the Networmonitor script…

HSD99 you are correct, Ripapa, apologies for the confusion

Cheers

C

Glad you’re back in business!

1 Like

Doesn’t make a lot of sense though. The modified Start Network Monitor script basically logged ‘Exiting Network Monitor’ and that’s it!

C

Correct, but you can’t just delete the file for example. It causes all kinds of problems. the vera still expects to execute it and get a return code!

For sure!

This was my Start_NetworkMonitor.sh:

#!/bin/sh

# SVN : $Id: Start_NetworkMonitor.sh 11683 2014-06-20 09:19:33Z florin $

#Copyright (C) 2008 Mi Casa Verde, Inc., a Nevada Corporation
#                    www.micasaverde.com
#           1 - 702 - 4879770 / 866 - 966 - casa
#This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License.
#This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
#without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

exit 0

log "Terminating Start_NetworkMonitor.sh"
log "===END==="

Which I thought I copied from you.

Cheers

C

Now I see what you mean… Well this has not caused any reload for me on any of my machines so I am not sure what is happening with yours… Do you have any plugins calling the internet on your vera maybe?

Entirely possible but it’s not like I don’t get LuuP restarts anyway. Maybe there’s something different at the start of them month

In other news I do have this error:

50103066_LuaUPnP.log_2019-07-03_09-49-03.gz:01 07/02/19 21:48:11.376 JobHandler_LuaUPnP::DownloadPlugin 9092 failed to convert error -4: ERROR: no version relased for this plugin <0x76fda520>
50103066_LuaUPnP.log_2019-07-03_09-49-03.gz:01 07/02/19 21:48:11.378 JobHandler_LuaUPnP::AutoInstallPlugins Error installing plugin 9092:(null) <0x76fda520>

Doesn’t seem to cause any issues, but would rather like it to stop. Only I can’t find out what plugin it is

But no idea how I find out what 9092 is…

I guess to find out if I’ve got anything calling out plugin wise, I’d have to do a tcp dump or similar?

Cheers

C

read here: https://community.getvera.com/t/beta-7-0-29-3-apr-logs-messages-for-missing-plugin-9092/207812/13

My thanks!

C

One new discovery on reloads: I tested overloading my zwave network using a secondary controller sending a lot of commands (it wasn’t all that much, I just sent NIF gets to about 20 devices) causing delays on the vera zwave serial api. It cause the luup engine to reload as well. Obviously the reload helped absolutely nothing but it is the built in flakiness of the engine. It seems even from the comments I read from the scripts in the firmware that the devs use the luup reload as their solution to everything to the point of being ridicule. (reminds me of windex of my big fat greek wedding movie). It is like giving an open heart surgery as a cure for an indigestion or a headache, not realizing that the luup reload itself is often the source of a lot of problems and is often much worse than the problem it is trying to resolve while failing to resolve it. At least on my system, it is the case for the vast majority of the events. The same can be said of the nightly heal: It’s overloading a perfectly healthy network on a regular basis causing all kinds of problems and luup reloads.

1 Like