| @g_gobi:matrix.org | Hi, | 09:30 |
|---|---|---|
| I tried to tun 5 jobs in a zuul pipeline. | ||
| of them were in runnning and remaining two are in pending state (waiting until closer to head of queue to start jobs) seeing this msg in console. Anyone know how to run multiple jobs at parallel? | 00:00 | |
| I'm using static nodepool. In nodepool's max-parallel-jobs set to 10. | ||
| @g_gobi:matrix.org | * Hi, | 09:31 |
| I tried to run 5 jobs in a zuul pipeline. | ||
| of them were in runnning and remaining two are in pending state (waiting until closer to head of queue to start jobs) seeing this msg in console. Anyone know how to run multiple jobs at parallel? | 00:00 | |
| I'm using static nodepool. In nodepool's max-parallel-jobs set to 10. | ||
| @g_gobi:matrix.org | * Hi, | 11:44 |
| I tried to run 5 jobs in a zuul pipeline. | ||
| of them were in running and remaining two are in pending state (waiting until closer to head of queue to start jobs) seeing this msg in console. Anyone know how to run multiple jobs at parallel? | 00:00 | |
| I'm using static nodepool. In nodepool's max-parallel-jobs set to 10. | ||
| @fungicide:matrix.org | > <@g_gobi:matrix.org> Hi, | 11:45 |
| > I tried to run 5 jobs in a zuul pipeline. | ||
| > 3 of them were in running and remaining two are in pending state (waiting until closer to head of queue to start jobs) seeing this msg in console. Anyone know how to run multiple jobs at parallel? | ||
| > | ||
| > I'm using static nodepool. In nodepool's max-parallel-jobs set to 10. | ||
| "waiting until closer to head of queue" is what zuul displays when a change is outside the active window. you can adjust the pipeline.window* options to tune that behavior: https://zuul-ci.org/docs/zuul/latest/config/pipeline.html#attr-pipeline.window | ||
| @g_gobi:matrix.org | fungi: looks like the default value is 20. I'm just trying to run 5 jobs. Is there any relation between these? | 11:47 |
| @g_gobi:matrix.org | * fungi: looks like the default value for **pipeline.window** is **20**. I'm just trying to run 5 jobs. Is there any relation between these? | 11:48 |
| @g_gobi:matrix.org | window-increase-factor(https://zuul-ci.org/docs/zuul/latest/config/pipeline.html#attr-pipeline.window-decrease-factor) what is the meaning of unsuccessful change merge? | 12:03 |
| @q:fricklercloud.de | are these 5 different changes with one job each, or 5 jobs for a single change? the latter shouldn't be affected by the window at all | 12:32 |
| @q:fricklercloud.de | and the decrease factor kicks in when runs in that pipeline have been failing, which would match the minimum of 3 you seem to be seeing | 12:33 |
| @g_gobi:matrix.org | 5 different MRs | 12:35 |
| @g_gobi:matrix.org | frickler: you mean because of previous mrs failed, window size should be shrink. So, it can't match the min value 3. Because of that it went pending? | 12:39 |
| If 3 jobs failed, | ||
| 1) 20/2=10 | ||
| 2) 10/2 = 5 | ||
| 3) 5/2 = 2.5 | ||
| so 4th job won't match min window requirement 3. So, it is pending. Am I right? | ||
| @q:fricklercloud.de | that's how it looks like to me, yes. | 12:42 |
| @g_gobi:matrix.org | * frickler: you mean because of previous mrs failed, window size should be shrink. So, it can't match the min value 3. Because of that it went pending? | 12:42 |
| default decrease factor =2 | ||
| If 3 jobs failed, | ||
| 1. 20/2=10 | ||
| 2. 10/2 = 5 | ||
| 3. 5/2 = 2.5 | ||
| so 4th job won't match min window requirement 3. So, it is pending. Am I right? | ||
| @g_gobi:matrix.org | Can we disable window by setting 0. Whether it will affect somewhere? | 12:43 |
| @q:fricklercloud.de | either set the window to 0 or increase window-floor to a value you are more comfortable with. in the end zuul might get unhappy if having to deal with 100s of events in the pipeline at the same time, but that depends on your deployment scenario | 12:48 |
| @g_gobi:matrix.org | Okay. Thanks for the details. | 12:53 |
| @g_gobi:matrix.org | Setting pipeline.window-decrease-type : linear and pipeline.window-decrease-factor : 1 (same as increase factor) should be a good choice then. | 12:58 |
| @jim:acmegating.com | Gobi G: or you can adjust the floor to whatever minimum you want. the setting is mostly to avoid wasting cpu time if your software fails tests often, since zuul will need to restart jobs in that case. | 13:07 |
| @g_gobi:matrix.org | corvus: got it. In my case pass:fail ratio will be 40:60. So, in my case every failed job will reduce window by half because default it was exponential. | 13:12 |
| @g_gobi:matrix.org | Once my tests were high success rate that time "adjust the floor to whatever minimum" it will be useful. | 13:13 |
| @g_gobi:matrix.org | * Once my tests will become high success rate that time "adjust the floor to whatever minimum" it will be useful. | 13:13 |
| -@gerrit:opendev.org- Jeremy Stanley https://matrix.to/#/@fungicide:matrix.org proposed: [zuul/zuul-jobs] 844291: DNM: exercise base-test job https://review.opendev.org/c/zuul/zuul-jobs/+/844291 | 13:14 | |
| -@gerrit:opendev.org- Jeremy Stanley https://matrix.to/#/@fungicide:matrix.org proposed: [zuul/nodepool] 844334: Temporarily pin OpenStackSDK before 0.99 https://review.opendev.org/c/zuul/nodepool/+/844334 | 20:27 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 843737: Add prune-database command https://review.opendev.org/c/zuul/zuul/+/843737 | 21:29 | |
| -@gerrit:opendev.org- Zuul merged on behalf of Jeremy Stanley https://matrix.to/#/@fungicide:matrix.org: [zuul/nodepool] 844334: Temporarily pin OpenStackSDK before 0.99 https://review.opendev.org/c/zuul/nodepool/+/844334 | 23:48 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!