search

lanl / BEE

Other
13 stars 3 forks source link

Restart GDB #759

Closed rstyd closed 1 month ago

rstyd commented 6 months ago

This PR addresses #704.

jtronge commented 6 months ago

I tried this with examples/clamr-ffmpeg-build. I killed neo4j while the clamr task was running and it looks like the WFM restarted neo4j. I think the clamr task completed but the state is still at RUNNING and it doesn't get to the ffmpeg step.

rstyd commented 6 months ago

Oh I'll check that out. Thanks! I bet the code that normally sets that is getting short skipped.

pagrubel commented 6 months ago

I tried this with examples/clamr-ffmpeg-build. I killed neo4j while the clamr task was running and it looks like the WFM restarted neo4j. I think the clamr task completed but the state is still at RUNNING and it doesn't get to the ffmpeg step.

Oh I'll check that out. Thanks! I bet the code that normally sets that is getting short skipped.

You may also need to look at what commands you are using to check the status. I noticed in the past if a job finishes the status might not be available on some systems.

pagrubel commented 3 months ago

@rstyd There have been several merges into develop so please merge develop in this branch

jtronge commented 2 months ago

This works for me. I tried the clamr-ffmpeg-build workflow and killed the gdb twice, during the clamr and then the ffmpeg step, and both times it was restarted.

@rstyd Can you merge/rebase on top of develop to check against the recent neo4j updates? I'll approve after that.

pagrubel commented 2 months ago

@rstyd I also tried this and it works for me. Would you please make a squash commit with the important description of what was done. There have been many changes to develop since you started.

rstyd commented 2 months ago

No problem!

pagrubel commented 2 months ago

Also, reexcute works well, thanks for fixing that!!!