mag1024 / bosch-alarm-homeassistant

Integration for Bosch Alarm panels that use the mode2 protocol: Solution, B/G series, AMAX
21 stars 3 forks source link

Solution 3000 going unavailable #34

Open mapota63 opened 4 months ago

mapota63 commented 4 months ago

Hi,

Bosch Solution 3000 firmware v2.1.3 B426-M firmware 3.13.0208

Connects OK but after an unspecified time period, ranging from only hours to days it becomes unavailable.

Only way to get it back is to remove power from the B426-M board.

Whilst HA shows unavailable, I am able to ping the board with replies, and the RSC+ app works OK, but unable to load the B426-M webpage and uable to conenct using the Alink software (v6.1.16)

Only log I see is the following (though debugging is not turned on.)

Logger: bosch_alarm_mode2.panel Source: runner.py:188 First occurred: 10:21:50 AM (2 occurrences) Last logged: 12:50:35 PM

Heartbeat expired (0:03:04.127621): resetting connection. Heartbeat expired (0:03:29.526957): resetting connection.

I currently have the integration removed to try and determine if it is even related to this integration.

I will try at least once per day to connect using the ALink software to see if it is any better without the integration.

If it is OK, without the integration, what debug settings should I invoke to assist in troubleshooting.

Thank you for the great work you have done so far.

Andrew Potter

sanjay900 commented 4 months ago

I had this problem and then i realised i had legacy TCP turned on which was what caused it, is that setting on for you?

mapota63 commented 4 months ago

Hi, Sanjay, thanks for the quick response.

I am pretty sure that is not selected.

From A-Link Plus

image

and

image

from B426 Webpage

image

Andrew

sanjay900 commented 4 months ago

Update your b426-m firmware, they are up to v3.13.208 https://commerce.boschsecurity.com/au/en/B426-Conettix-Ethernet-Communication-Modules/p/F.01U.332.307/

Older B426-M firmwares are known to have issues like this

DataBitz commented 4 months ago

As you are running 2.1.3 firmware, you can reset the network module from the alarm control pad. It would be interesting to see if that gets it back. Enter your Master Code, followed by the [9][4][1] and [#] keys to reset the network module 1 https://resources-boschsecurity-cdn.azureedge.net/public/documents/RLN_V02_Release_Note_enUS_97508780683.pdf

sanjay900 commented 4 months ago

I used to get this exact problem. Resetting the network module works, but its only a temporary fix.

The network modules local connection gets locked up, but the cloud connection still works, hence why A-Link stops working as well but RSC+ continues working. A-Link actually still works if you use cloud instead of ip.

sanjay900 commented 4 months ago

One other thing bosch mentioned is that you can set the following settings and with the latest firmware it should restart the module automatically when it locks up like this

"Cell Modem Reset Count =10" (set Network module programming>>Module 1#/2#>>Other options>>Cell Modem Reset Count =10)

Make that change on both modules, even if you aren't using both of them

mapota63 commented 4 months ago

Thanks Sanjay. I am on v3.13.208 already. I am going to leave the integration off for a few more days just to prove the integration is at fault, even if it just the more frequent traffic. Once that time has elapsed, likely early next week. I will try resetting via panel and then add the auto resetting via the panel config.

image
jayvarads commented 4 months ago

Hi Andrew,

My test setup has been connected for a month without disconnection. Only difference I can see you have DHCP enabled.

Bosch Solution 3000 firmware v2.1.3 B426-M firmware 3.13.0208

Have you tried to set a manual IP address for the B426-M in A-link plus?

image

mapota63 commented 4 months ago

I have not used a manual Ip as yet will also try that. I did have my router set to always use the same IP

Andrew

From: jayvarads @.> Sent: Thursday, March 14, 2024 8:22 AM To: mag1024/bosch-alarm-homeassistant @.> Cc: mapota63 @.>; Author @.> Subject: Re: [mag1024/bosch-alarm-homeassistant] Solution 3000 going unavailable (Issue #34)

Hi Andrew,

My test setup has been connected for a month without disconnection. Only difference I can see you have DHCP enabled.

Bosch Solution 3000 firmware v2.1.3 B426-M firmware 3.13.0208

Have you tried to set a manual IP address for the B426-M in A-link plus?

image.png (view on web) https://github.com/mag1024/bosch-alarm-homeassistant/assets/112724328/368d6d80-0cad-4fbe-b01d-e142c2cc9e09

— Reply to this email directly, view it on GitHub https://github.com/mag1024/bosch-alarm-homeassistant/issues/34#issuecomment-1995867891 , or unsubscribe https://github.com/notifications/unsubscribe-auth/AI7ALAYRA7JUVIBZKVMEKLDYYC7QXAVCNFSM6AAAAABERNSV2CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOJVHA3DOOBZGE . You are receiving this because you authored the thread. https://github.com/notifications/beacon/AI7ALA3LY6OLBO3GTQX3LFLYYC7QXA5CNFSM6AAAAABERNSV2CWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTTW62DPG.gif Message ID: @. @.> >

jayvarads commented 4 months ago

I have not used a manual Ip as yet will also try that. I did have my router set to always use the same IP Andrew From: jayvarads @.> Sent: Thursday, March 14, 2024 8:22 AM To: mag1024/bosch-alarm-homeassistant @.> Cc: mapota63 @.>; Author @.> Subject: Re: [mag1024/bosch-alarm-homeassistant] Solution 3000 going unavailable (Issue #34) Hi Andrew, My test setup has been connected for a month without disconnection. Only difference I can see you have DHCP enabled. Bosch Solution 3000 firmware v2.1.3 B426-M firmware 3.13.0208 Have you tried to set a manual IP address for the B426-M in A-link plus? image.png (view on web) https://github.com/mag1024/bosch-alarm-homeassistant/assets/112724328/368d6d80-0cad-4fbe-b01d-e142c2cc9e09 — Reply to this email directly, view it on GitHub <#34 (comment)> , or unsubscribe https://github.com/notifications/unsubscribe-auth/AI7ALAYRA7JUVIBZKVMEKLDYYC7QXAVCNFSM6AAAAABERNSV2CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSOJVHA3DOOBZGE . You are receiving this because you authored the thread. https://github.com/notifications/beacon/AI7ALA3LY6OLBO3GTQX3LFLYYC7QXA5CNFSM6AAAAABERNSV2CWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTTW62DPG.gif Message ID: @. @.> >

It could be the B426-M trying to renew IP lease even though the router has the static IP reservation. Best to put static IP in the B426-M module. This is what we found to work best.

sanjay900 commented 4 months ago

Is your network unstable by any chance? I just had the same thing happen to my panel after it working fine for several months, and noted that a bunch of devices on my network had reconnections at that point, so there was just a blip where my network stopped working.

Resetting via the panel does work, but i knew that from previous experience.

mapota63 commented 4 months ago

Will keep that in mind. I have added a raspberry pi with wireshark to check network traffic. Once i add the integration back on I will set it up.The system has been ok without the integration. I.e i can connect with web page and alink plus whenever I like without needing to reset.I am consolidating my ip addresses to have more static ips ( ready to set the alarm as static ) as suggested by jayvarads.

mapota63 commented 4 months ago

OK,

Have set a static IP for the alarm Module, rather than DHCP.

Have also reinstalled the integration, will see how it goes. Wireshark is capturing packets

Andrew

From: Sanjay Govind @.> Sent: Tuesday, March 19, 2024 8:41 AM To: mag1024/bosch-alarm-homeassistant @.> Cc: mapota63 @.>; Author @.> Subject: Re: [mag1024/bosch-alarm-homeassistant] Solution 3000 going unavailable (Issue #34)

Is your network unstable by any chance? I just had the same thing happen to my panel after it working fine for several months, and noted that a bunch of devices on my network had reconnections at that point, so there was just a blip where my network stopped working.

Resetting via the panel does work, but i knew that from previous experience.

— Reply to this email directly, view it on GitHub https://github.com/mag1024/bosch-alarm-homeassistant/issues/34#issuecomment-2005078665 , or unsubscribe https://github.com/notifications/unsubscribe-auth/AI7ALA7AZUQWWS5HRGPNKZ3YY5NODAVCNFSM6AAAAABERNSV2CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMBVGA3TQNRWGU . You are receiving this because you authored the thread. https://github.com/notifications/beacon/AI7ALA5K3DJZ3AP4TFCUXQTYY5NODA5CNFSM6AAAAABERNSV2CWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTTXQMJIS.gif Message ID: @. @.> >

mapota63 commented 4 months ago

OK,

Still seeing issues. I do see a lot of TCP retransmissions in the wireshark captures. Trying to get to the bottom of it. I know some are expected but I seem to be seeing too many. Will keep checking.

One thing I have noticed is when I reset the ip module, i need to reload the integration (restarting HA also works, as it would reload the integration as well). But when it comes back up, the alarm panel now has a code requirement.

If I remove the integration and reinstall then the requirement for a code disappears.

image

How do I prevent that?

Andrew

sanjay900 commented 4 months ago

OK,

Still seeing issues. I do see a lot of TCP retransmissions in the wireshark captures. Trying to get to the bottom of it. I know some are expected but I seem to be seeing too many. Will keep checking.

One thing I have noticed is when I reset the ip module, i need to reload the integration (restarting HA also works, as it would reload the integration as well). But when it comes back up, the alarm panel now has a code requirement.

If I remove the integration and reinstall then the requirement for a code disappears. image

How do I prevent that?

Andrew

There's a setting for the code requirement within the integration, it shouldn't be enabling itself though that's a bit bizarre

If you go to the settings at that point what do you see?

mapota63 commented 4 months ago

I am wondering if I have the latest integration. I have looked in the integration but could not see a switch, either on or off. Obviously not looking in the right place though.Get Outlook for Android

mapota63 commented 4 months ago

Ok found it, I incorrectly assumed it was asking me for the  master code as per when the integration is first installed, so I entered that, did not realise that if I left it blank the code buttons do not appear. So all my fault. Sorry.Get Outlook for Android

sanjay900 commented 4 months ago

No problem at least that was an easy fix

mapota63 commented 2 months ago

Appears to be network related, but still trying to figure it out exactly.

I did have a rasperrry pi4 runnnig as a zabbix server. When that is up and running the alarm goes unavailable often. Tried shutting it down and alarm works ok for three or four days, restarted zabbix server and alarm only stayed available for about 4 hours then 5 hours. Reshutdown zabbix server and was OK for about 12 days.

Became unavailable at 14:34:18 on the 8th May.

Looking at HA entity history Note panel was reset using the keypad (941 code) at around 16:33

for Garage Doors I can see 08-05-2024 14:29:14 Became Open 08-05-2024 14:29:14 Became Open 08-05-2024 14:34:18 Became Unavailable 08-05-2024 16:32:37 Became Closed

For Armed Status I can see 08-05-2024 14:31:16 Changed to Arming Triggered - Arm Away 08-05-2024 14:34:18 Changed to Unknown 08-05-2024 16:32:36 Changed to Disarmed 08-05-2024 16:32:37 Changed to unavailable 08-05-2024 16:33:01 Changed to Disarmed

From HA logs 2024-05-08 14:34:18.121 WARNING (MainThread) [bosch_alarm_mode2.panel] Heartbeat expired (0:03:02.063921): resetting connection. 2024-05-08 14:34:58.859 ERROR (MainThread) [root] Connection monitor exception Traceback (most recent call last): File "/usr/local/lib/python3.12/site-packages/bosch_alarm_mode2/panel.py", line 304, in _monitor_connection await self._monitor_connection_once() File "/usr/local/lib/python3.12/site-packages/bosch_alarm_mode2/panel.py", line 331, in _monitor_connection_once await self._connect(load_selector) File "/usr/local/lib/python3.12/site-packages/bosch_alarm_mode2/panel.py", line 257, in _connect await self._authenticate() File "/usr/local/lib/python3.12/site-packages/bosch_alarm_mode2/panel.py", line 400, in _authenticate await self._authenticate_remote_user() File "/usr/local/lib/python3.12/site-packages/bosch_alarm_mode2/panel.py", line 346, in _authenticate_remote_user raise PermissionError("'Master code functions' authority required") PermissionError: 'Master code functions' authority required 2024-05-08 14:37:42.580 ERROR (MainThread) [homeassistant.components.ipp.coordinator] Error fetching ipp data: Invalid response from API: Timeout occurred while connecting to IPP server.

In wireshark I can see the panel is talking to amazon AWS (assume that is the RSC+ app??) and to the local gateway all through the "unavailable" status.

I did notice some upnp stuff going on so have disabled upnp on my router (fritzbox). Wil see how it goes again.

As of now it is still connected and talking to HA.

The auto reset you mentioned in a early post does not seem to help, i.e. it does not auto reset! _"Cell Modem Reset Count =10" (set Network module programming>>Module 1#/2#>>Other options>>Cell Modem Reset Count =10)

Originally my value was a 5 !!

Andrew

sanjay900 commented 2 months ago

I am trying to work with bosch on this one since I've had it happen to me a few times as well, seeing if i can get some debug logs from my B426-M and work on it that way.

mapota63 commented 2 months ago

Sanjay,

Excellent news. Thanks for that, if you would like me to try any debugging let me know. I will be happy to do what I can.

Andrew

From: Sanjay Govind @.> Sent: Friday, May 17, 2024 8:58 AM To: mag1024/bosch-alarm-homeassistant @.> Cc: mapota63 @.>; Author @.> Subject: Re: [mag1024/bosch-alarm-homeassistant] Solution 3000 going unavailable (Issue #34)

I am trying to work with bosch on this one since I've had it happen to me a few times as well, seeing if i can get some debug logs from my B426-M and work on it that way.

— Reply to this email directly, view it on GitHub https://github.com/mag1024/bosch-alarm-homeassistant/issues/34#issuecomment-2116332406 , or unsubscribe https://github.com/notifications/unsubscribe-auth/AI7ALA75CBEDDEHIX7F74ADZCU2WVAVCNFSM6AAAAABERNSV2CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMJWGMZTENBQGY . You are receiving this because you authored the thread. https://github.com/notifications/beacon/AI7ALA2CDCJMJH2FVAIAA33ZCU2WVA5CNFSM6AAAAABERNSV2CWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTT6ESVXM.gif Message ID: @. @.> >

bdleung commented 2 months ago

Happening to me too. I'm finding when the integration is showing disconnected, the B426 is still responding to pings.

I had to restart the B426 (I have an esphome-enabled relay on the power feed to the B426) and then restart HA for it to be detected again. The B426 is responding to pings the whole time (except when power is off, of course).

I use DHCP with a reservation for the B426.

sanjay900 commented 2 months ago

Yeah from what i found the B426 locks up but only for local network connections - for whatever reason cloud connections still work which is weird. I have a debug enabled B426-M but have not been able to get it to work yet, still working on that one

mapota63 commented 2 months ago

Can confirm mine responds to pings but the A-Link software refuses to connect. Rsc+ link does seem to continue to work, as Sanjay noted.Get Outlook for AndroidFrom: Sanjay Govind @.>Sent: Sunday, May 19, 2024 8:38:46 pmTo: mag1024/bosch-alarm-homeassistant @.>Cc: mapota63 @.>; Author @.>Subject: Re: [mag1024/bosch-alarm-homeassistant] Solution 3000 going unavailable (Issue #34) Yeah from what i found the B426 locks up but only for local network connections - for whatever reason cloud connections still work which is weird. I have a debug enabled B426-M but have not been able to get it to work yet, still working on that one

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: @.***>

sanjay900 commented 2 months ago

I found you could easily get the panel into this state just by connecting to it with multiple copies of bosch-alarm-mode2 at once, and then got my debug logs working and sent a copy of the logs to bosch along with those replication steps, so hopefully they get to the bottom of this.

I'm also currently trying to see if i can get my test panel to drop off naturally as well, we will see what happens there. If it does stop working then ill send those logs to bosch as well, but as you know it can take multiple days for that to happen :(

DataBitz commented 2 months ago

My 2 cents. I've been running this for about 6 months, I only had the B426-M stop accepting local network connections about 4-5 times. Sometimes after Home Assistant updates or restarts, other times when I was using the A-Link or B426-M web interface (both I rarely use). No problems for a couple of months now.

sanjay900 commented 2 months ago

I was told some changes have been made to the RSC+ cloud servers, be interesting to see if that makes a difference or not.

lucidicious commented 4 weeks ago

Sanjay, I was experiencing regular connection dropouts immediately after arming my Solution 3000. I noticed this happening after I reduced the exit delay without taking into account the actual time it takes for my garage roller door to close, resulting in an isolated zone. Once I increased the delay time, I've haven't had a drop out. Not sure if such an event being flagged whilst armed could affect the integration. FWIW, it might be an example of why connections drop from time to time. I'm not using RSC+.

an0nym0us000 commented 4 weeks ago

I'm still getting issues with my Bosch 3000 going unavailable. Can't pinpoint what's causing it but definitely has been times where I've restarted home assistant and it has gone from connected to unavailable. It doesn't always do this but definitely has done so at a restart. Other times it appears to be working fine and then all of the sudden go to unavailable without me logging into the rsc+ app, restarting home assistant etc.

sanjay900 commented 3 weeks ago

Sanjay, I was experiencing regular connection dropouts immediately after arming my Solution 3000. I noticed this happening after I reduced the exit delay without taking into account the actual time it takes for my garage roller door to close, resulting in an isolated zone. Once I increased the delay time, I've haven't had a drop out. Not sure if such an event being flagged whilst armed could affect the integration. FWIW, it might be an example of why connections drop from time to time. I'm not using RSC+.

Interesting, unfortuntely I probably won't be able to investigate this for a few months though since im going on holiday soon

pauljmccarthy commented 3 weeks ago

Ok found it, I incorrectly assumed it was asking me for the  master code as per when the integration is first installed, so I entered that, did not realise that if I left it blank the code buttons do not appear. So all my fault. Sorry.Get Outlook for Android

I had this issue too with Solution 3000. On configuring the integration, was prompted for a code so I entered one. Whenever the alarm was armed, the entities would go unavailable. All fixed when removing the code from the configuration. Would be helpful to update the documentation on when to enter a code and when it should not be provided in the configuration.