CFE-3625: Increase Filedescriptor Limit to a more contemporary Value #4609
Conversation
Ticket: CFE-3625 Changelog: Title
flynn1973
changed the title
|
@cf-bottom please test this in our Jenkins CI system :) |
|
Alright, I triggered a build: Jenkins: https://ci.cfengine.com/job/pr-pipeline/6470/ Packages: http://buildcache.cfengine.com/packages/testing-pr/jenkins-pr-pipeline-6470/ |
vpodzime
left a comment
vpodzime
left a comment
There was a problem hiding this comment.
Looks good to me otherwise.
libpromises/pipes_unix.c
Outdated
|
|
||
| static pid_t *CHILDREN = NULL; /* GLOBAL_X */ | ||
| static int MAX_FD = 128; /* GLOBAL_X */ /* Max number of simultaneous pipes */ | ||
| static int MAX_FD = 8192; /* GLOBAL_X */ /* Max number of simultaneous pipes */ |
There was a problem hiding this comment.
Wow, big hammer approach. :) I'm not quite sure 8192 child processes (or even 4096) running at the same time is feasible. What about 2048?
There was a problem hiding this comment.
:) yeah, maybe a little bit overstated, i like to live dangerously. 2048 should be enough. i am trapped in AIX and big machine thinking.
There was a problem hiding this comment.
I was going to suggest 640k
There was a problem hiding this comment.
hmm...basically i can patch this locally, as i am constantly building cfengine rpms for aix. NorthernTechHQ/libntech#130 looks very "promising" to me, and i dont want to fire up a "numbers" flamewar :-). this problem seems to be a isolated thing (only me) anyway. nobody is doing stupid stuff like me g
There was a problem hiding this comment.
128 is definitely a very low and totally artificial limit. 2048 should be fine for everybody. 😉
|
Just a note: I'm currently working on NorthernTechHQ/libntech#130 which should be a better long-term solution for this and many other similar problems. Having said that, we should definitely make this simple change first. |
Co-authored-by: Ole Herman Schumacher Elgesem <olehermanse@users.noreply.github.com>
|
@flynn1973 could you please squash the commits? Thanks! |
No changelog entry since this memory is not really leaking, it's just not cleaned up properly right before an exit(). Changelog: None Ticket: CFE-3431 Signed-off-by: Ole Herman Schumacher Elgesem <ole@northern.tech>
Also added acceptance test. Changelog: unless can now be used with custom promise types Ticket: CFE-3431 Signed-off-by: Ole Herman Schumacher Elgesem <ole@northern.tech>
Ticket: CFE-3625 Changelog: Title
flynn1973
deleted the
CFE-3625_File_descriptor_of_child_higher_than_MAX_FD
branch
|
superseeded by #4615 |
5 participants
Ticket: CFE-3625
Changelog: Title