CPU collection and FS issues

[antoniogi]

CPU is not being reported on ST2.
If the filesystem is too busy, REMORA times out and leaves the files in the root output folder. This needs to be fixed.

[carlosrosales]

Run remora_post_crash at the end if one has to wait the full amount specified in the loop for the fs?

…

On Jun 23, 2017 1:02 PM, "Antonio Gómez" ***@***.***> wrote: CPU is not being reported on ST2. If the filesystem is too busy, REMORA times out and leaves the files in the root output folder. This needs to be fixed. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#39>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACGrmWKsh9oRZZTg1-cqSjbcxQCH7oc1ks5sG_2wgaJpZM4OD32S> .

[antoniogi]

That works, I think. But I'll move all the functionality of remora_post_crash into a function. Don't want to start calling scripts from source again :)

[carlosrosales]

Yeah, good point...

…

On Jun 23, 2017 1:23 PM, "Antonio Gómez" ***@***.***> wrote: That works, I think. But I'll move all the functionality of remora_post_crash into a function. Don't want to start calling scripts from source again :) — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#39 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACGrmXJ7ll35a__Cv0zZ2I4vKXJGoAGCks5sHAK4gaJpZM4OD32S> .

[antoniogi]

CPU problem fixed. When REMORA tries to use the mpstat binary that it's installed with the tool, it fails because the PATH in the remote nodes might not be set properly.
Working on the FS part now, that's not fun.

[antoniogi]

Calling remora_post_crash is still not solving all the issues when the FS is extremely slow. I have temporarily increased the timeout so that at least the problems are minimized. I will have to implement something more sophisticated for this later on.