Vera extroot

[Catman]

@scyto

Does it have to be an SSD or can use large USB stick ok?

@HSD99

They are slower and you’d probably notice under some circumstances. Honestly, using a real SSD with wear leveling and garbage collection is the way to go. I’ve been extrooted with a high-quality 128GB (what I had on-hand) SSD since September of 2018 and the system has performed flawlessly. The cost of a small SSD is not high.

Make sense?

C

[Catman]

So I’ll post this here, but not sure it’s completely the correct place. Let’s see.

This evening has been interesting. Started when I noticed that two of my Fibaro connected temp sensors had dropped out. I figured I’d do a restore from just before they failed.
So I did just that thing. Only to be met with a dead box.
All the lights are on.
SSH works
Odd
Reboot. Power comes on. Then nothing. More nothing, then all the lights come on at the same time.
Odd still.
SSH, no interface. Controller is off line
Much discussion with one of the guys on the helpdesk and we unplug the extroot SSD
Controller comes back as it was in February (just before the extroot)
Long story short, every time I plug the SSD back in, all sorts of weird shit happens. Including the LAN light on the connected switch going off…
SSD is fine in my Mac, so re-initialised. Still makes all the lights go mad.

I can only assume it’s failed in some way. Any thoughts?

Cheers

C

[blacey]

@Catman

I can only assume it’s failed in some way. Any thoughts?

SSDs can be power hungry so perhaps the one you are using exceeds the Vera’s available USB current. If so, then the SSD will not operate correctly and cause USB port interruptions. Try attaching the SSD to a USB powered hub connected to the Vera or a lower-power SSD.

[rafale77]

SSH still worked? This is very strange. If the vera failed to boot, then I would have suspected a power issue but it seems to have booted fine. If upon reboot it did not come back, then it is likely that somehow the extroot itself failed. But the first failure makes me suspect an error in the USB communication which may in turn have caused a corruption of the fstab. If you could post your fstab config, maybe I could take a look.

[Catman]

@blacey

Try attaching the SSD to a USB powered hub connected to the Vera or a lower-power SSD.

This is a good shout, although it’s been fine for a couple of months.

@rafale77

SSH still worked? This is very strange. If the vera failed to boot, then I would have suspected a power issue but it seems to have booted fine. If upon reboot it did not come back, then it is likely that somehow the extroot itself failed. But the first failure makes me suspect an error in the USB communication which may in turn have caused a corruption of the fstab. If you could post your fstab config, maybe I could take a look.

Lua would not start. I’ve going to try some other USB devices and controllers see if it is the SSD or the controller for that.

But I think I may be in a bit of a halfway house and not sure how to get out. I seem to be missing lots of log files, for example I only have three logs in /var/log/cmh (LuauPnP, InternetOK and NetworkMonitor.

When we did a factory reset with the SSD in, it didn’t work. No power light, but everything else on. Hard reboot brought it back to SSH only.
Removing the SSD and power cycling again took it back to what looks like a Feb set up, just before I extrooted it.

I’ve restored back to late March (when all my sensors were working) but they’re still not and as I say, I suspect I’ve lost some files in there somewhere.

I guess the sensible thing now:

1. Identify if the SSH, the SSD cable of the Plus USB is faulty.
2. Factory reset nakedVera
3. Restore from pre extroot backup.
4. Extroot

Or is there a simpler way?

TIA!

C

[Catman]

OK just plugged in a USB stick and re-enabled USB logging. /var/log/cmh looks rather more healthy now, but still not sure it’s complete?
-rw-r–r-- 1 root root 257 Apr 8 18:11 InternetOk.log
-rw-r–r-- 1 root root 241804 Apr 8 18:18 LuaUPnP.log
-rw-r–r-- 1 root root 45001 Apr 8 18:19 NetworkMonitor.log
drwx------ 2 root root 12288 Apr 8 18:09 lost+found
-rw-r–r-- 1 root root 2128 Apr 8 18:11 serproxy.log

Looks like it may well be the SSD or the cable. That’ll be fun checking
C

[Catman]

Talking to myself and the future Feeling relaxed turned off the USB logging and plugged the SSD back in on the same controller cable.
Boom
Waited for everything to come back and then used a different cable.
All is well and I’m back extrooted.
So odd things are afoot with my other controller cable. Still works fine with a laptop, though. Shrug.
For some reason my logrotation and servers have not been restored but there we are

C

[richie.digital]

How do we increase ram file after extroot ?

[rafale77]

The vera has no ram file…

[richie.digital]

@rafale77

End note. You can get creative and do all kinds of things with this. For example add a swap file so the vera can have more available RAM etc…

^^ this is what im talking about

was it this

OpenWRT – Rootfs & Swap on USB Storage

[rafale77]

Ohh… I see yeah you can create a “swapfile”. It is pretty simple to do. You can go to the fstab and get it started from there. I have not had to do since I don’t have any RAM problem.

https://forum.archive.openwrt.org/viewtopic.php?id=17002

[richie.digital]

not sure if i have a memory problem here but even though my vera doent reboot the luup engine restarts

i have a lot of apps running

[rafale77]

This looks like you have a memory leak and a rather fast one. I would address it first. Adding more memory is only going to delay the crash, not eliminate it. Unless you see it stabilize at some point… I am not seeing it from your data.

[richie.digital]

@rafale77

h, not eliminate it. Unless you see it stabilize at some poin

great point … where do i start with this memory leak thing:slightly_smiling_face:

[rafale77]

The brute force way is to either:

1. remove all your plugins and reinstall them one at a time making sure you don’t see this abrupt downward slope in available memory. (assuming the blue line is the available memory)
2. Remove plugins one at a time.

Unfortunately, since the plugins all run within the LuaUPnP process, there is not other way to break them down and see how much each one is using at the OS level.

What is the red line? The sawtooth pattern suggests that it is related to the log rotate function and you can see that the frequency of the rotation increases over time. Is it the free memory?

[rigpapa]

It may be useful if you posted a list of the plugins you have installed. Some may be recognized as troublesome. There’s lots of history and tribal knowledge there.

[richie.digital]

Global Cache GC100

Details

Virtual Outdoor Temperature Plugin (by MiOS)

Details

Amazon Alexa Helper

Details

Virtual ON/OFF Switches

Details

Push Notification

Details

Blue Iris Camera Plugin

Details

HouseModes Plugin

Details

System Monitor

Details

OpenSprinkler

Details

SiteSensor

Sercomm IP Camera

Details

Broadlink Vera Concierge

Details

Virtual Clock

Details

MultiSwitch

Details

Combination Switch

Details

Denon Receiver Control Plugin

Details

Garage Door

Details

Philips Hue ALTHue plugin

Details

VeraAlerts

Details

Kodi Remote

Details

Day or Night

Details

Reactor

Details

Radio Thermostat Wi-Fi Plugin

Details

VeraSecure

Details

Ping Sensor

Details

LuaView

Details

AutoVera

Details

ALTUI

Details

ERGY Plugin

Details

Switchboard

Details

Wake-on-LAN

Details

[rigpapa]

Well, I believe the Ergy plugin has a long history of suspicion. See thread Ergy? - Power Management - Ezlo Community

[HSD99]

Agree 100%. I removed it from my system years ago due to a variety of problems.

[rafale77]

Wow, that’s a long list. Some have very similar functionalities… Are you stress testing the vera? Ergy would be my prime suspect as well. Not sure what the VeraSecure is. Are you making use of all of them?