I get the following error occasionally which I need to trap and issue something like “sudo systemctl restart nodered” to restart node-red automatically. Please share if you know how!!
Error connecting to dbus: Error [ERR_STREAM_DESTROYED]: Cannot call write after a stream was destroyed
Can i assume node-red is being crashed and the error in the log is ERR_STREAM_DESTROYED?
If so, you should identify the node causing this and report that as an issue on the nodes repository (in short, something is possibly not handing errors correctly & can probably be fixed to not crash node-red)
As for restarting node-red service - surely you just need to setup auto restart?
Yes, that is correct. I have identified the node where it is caused but getting no response from an earlier issue. As a VERY new user to this sort of thing (node-red is the limit of any "programming" I have done) the auto restart is the sort of thing I need but not sure how to add it to node-red.
If node-red is crashing, it cannot restart itself.
However if you mean you are trying to get node-red to setup auto restart, then you wouldn't do that either. It is systemctl itself that takes care of auto restarting a failed service.
Re-read that link I provided in my first response or search the internet for "systemctl auto restart failed service"
the node is @victronenergy/node-red-contrib-victron.
i'm not sure if it is node red crashing or just stopping as the erroer i get is just displayed as "Error connecting to dbus: Error [ERR_STREAM_DESTROYED]: Cannot call write after a stream was destroyed". What i am trying to replicate is the manual action of typing node-red-stop and node-red-start in a terminal window when it happens, a big issue being I dont know it has happened until i look
Seems the developers are aware of this issue. Seems to me they are not so interested in fixing it as they dont see this problem on thier own OS! I would (if I were you) post a follow up reply in that issue stating you too are having the issue - just to let the developer know the issue still exists & is probably affecting more people than they realise (they may not be aware of how common this is as people tend not to bother reporting issues and generate workarounds - much like yourself)
In the mean time, there are several threads on this forum related to restarting node-red this oneMIGHT work for you (it restarts your flows via the API - import the last flow under "solution 2" in that link)