Bug 12090

Summary: bitbake memory resident server reconnect handling needed
Product: [Build System, Metadata & Runtime] BitBake Reporter: brian avery <brian.avery>
Component: bitbakeAssignee: Richard Purdie <richard.purdie>
Status: NEW --- QA Contact:
Severity: normal    
Priority: Medium+ CC: bluelightning, poky.bs.watcher, poky.watcher, randy.macleod
Version: 2.4   
Target Milestone: 6.99   
Hardware: x86   
OS: Multiple   
Whiteboard:
OS type for building Yocto: --- Type of Regression: ---
Verified: Documentation change: Don't know
Bug Depends on:    
Bug Blocks: 13023    

Description brian avery 2017-09-14 22:42:24 UTC
The following set of actions can be an issue:

1) from a laptop, ssh into a build server and run:
$ cd ~/cow
$ . ./poky/oe-init-build-env
$ bitbake core-image-sato

Wait until the server is into something that takes a long time like building glibc

Then close laptop, or disconnect it from the network.

2) From a terminal on the server itself
$ cd ~/cow
$ . ./poky/oe-init-build-env
$ bitbake core-image-sato

This results in 
20 seconds of *no output* !!! 
Followed by:
Reconnecting to bitbake server ....
Retrying server connection ....
a number of times followed by:
ERROR: Unable to connect to bitbake server or start one

This is because the original bitbake server  (whose pid is in bitbake.lock) is still merrily building the image.  unfortunately, there isn't afaik, a way to reconnect to the running server, nor is it terribly obvious that the reason you can't run another bitbake command is that there is a running server already using that build directory.

+++++
Furthermore, Paul pointed out that if you did enable a reconnection, you might need to add a new event (welcome back?) that would fire on reconnection so that the connection didn't look dead in case the server was busy working but had no events to hand out.
Comment 1 Randy MacLeod 2023-10-24 14:24:29 UTC
Bulk move from 4.99 or 0.00 to 5.99