Second lyse instance

[philipstarkey]

Original report (archived issue) by Jan Werkmann (Bitbucket: PhyNerd, GitHub: PhyNerd).

---

It would be nice to have to option to open more then one instance of lyse. That way one can do some analysis in one instance with old data while new data is processed in in the other.

It's not hard to deactivate the server on a second instance of lyse with a simple try except and lyse will as often as one wants. However this causes problems since the dataframe is being transmitted with that server. So the second instance of lyse would still need a server.

We could add a way to communicate a new port to the analysis processes so that they to the second server.

[philipstarkey]

Original comment by Philip Starkey (Bitbucket: pstarkey, GitHub: philipstarkey).

---

I'm personally in favour of just having a second lyse computer (or virtual machine) if you want to run two instances

[philipstarkey]

Original comment by Chris Billington (Bitbucket: cbillington, GitHub: chrisjbillington).

---

My preferred approach here, if you wanted a single computer, would be to allow a command line argument to specify the config file to be used. I think the configparser library might allow for configurations to be hierarchical with respect to files, such that you could specify a config file that only included a few values, and anything not specified would fall back to the default labconfig. Then you could just specify a config file that had a different port number. The config file would have to be communicated to child processes so that they would get the right port for their interactions with the lyse server.

[philipstarkey]

Original comment by Jan Werkmann (Bitbucket: PhyNerd, GitHub: PhyNerd).

---

We are currently using a second computer to achieve this but the one computer solution seems superior to me.

I'll look into the config file approach that shouldn't be to hard to implement if hierarchical configs work.