Fix Connection is being forcefully terminated error in GUI

[andrei-toterman]

we have bidirectional streaming rpcs, but we usually send only one response back on that stream for most operations. therefore, the gui would only read data from the stream until it got the first reply. but this introduced errors in which the stream was disposed too early. now we read everything from the stream, so that issue shouldn't happen.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.95%. Comparing base (7074116) to head (7f7d7fd).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3829   +/-   ##
=======================================
  Coverage   88.94%   88.95%           
=======================================
  Files         256      256           
  Lines       14584    14584           
=======================================
+ Hits        12972    12973    +1     
+ Misses       1612     1611    -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Sploder12]

@andrei-toterman, could you share repro steps for the issue this solves?

[andrei-toterman]

@Sploder12 The problem is that it doesn't happen all the time. But you can try having two stopped instances, go to the instance list page, start one instance and then immediately try starting the other one. That was one way I managed to reproduce it a few times. The other way I could reproduce it all the time is with the new 'Open shell' button functionality, which would start the instance if its stopped. But that PR has this branch as a base, precisely because the issue would occur all the time. So if you want a good way to test it, you could rebase that branch on main, see if the issue occurs, than rebase it back on this branch and see the issue not occur anymore. Sorry for not having better steps, it's kind of an elusive issue 😅

[Sploder12]

@Sploder12 The problem is that it doesn't happen all the time. But you can try having two stopped instances, go to the instance list page, start one instance and then immediately try starting the other one. That was one way I managed to reproduce it a few times. The other way I could reproduce it all the time is with the new 'Open shell' button functionality, which would start the instance if its stopped. But that PR has this branch as a base, precisely because the issue would occur all the time. So if you want a good way to test it, you could rebase that branch on main, see if the issue occurs, than rebase it back on this branch and see the issue not occur anymore. Sorry for not having better steps, it's kind of an elusive issue 😅

Ah, I see, thanks!

[Sploder12]

LGTM! Just a small DRY comment

[Sploder12]

LGTM! Nice job!

[andrei-toterman]

@Sploder12 oops, the behavior after my refactor was not consistent with the previous behavior. Before, we were not logging all of the requests and replies for the info rpc, because we poll it frequently, so it would've been too much logging. But with the refactor, now there were constant logs for all the sent info requests. Now it should be fixed.

[Sploder12]

Oops, I missed that too. I paid extra attention to the logging this time and it looks good to me!