| Summary: | Add in tests for re-execution of tasks | ||
|---|---|---|---|
| Product: | [QA/Testing] Test Plans/Suite | Reporter: | Richard Purdie <richard.purdie> |
| Component: | test-plans/suite | Assignee: | Jiajun Xu <jiajun.xu> |
| Status: | VERIFIED FIXED | QA Contact: | |
| Severity: | enhancement | ||
| Priority: | Low | CC: | jessica.zhang, laurentiu.serban, sanil.kumar, song.liu, venkatarg, yp.test.watcher, yp.watcher |
| Version: | unspecified | ||
| Target Milestone: | 1.3 M1 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | (P2)(No QA needed)(patch checked in) | ||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | --- | |
| Attachments: | |||
|
Description
Richard Purdie
2012-03-15 13:37:45 UTC
I tried master with commit (82b6146ef29f06b....), there are 13 tasks failed due to different reasons against qemux86 core-image-sato-sdk. Attach the failure log and re-run log for reference. Created attachment 401 [details]
13 failed tasks with force against qemux86 core-image-sato-sdk
Jiajun, do you have the script you used to generate these results? Created attachment 402 [details]
script to force run all tasks
(In reply to comment #3) > Jiajun, do you have the script you used to generate these results? Yes. I upload the script here. Could you help to check if it's OK? I would like to add it into 1.2 M4 testing, and mark the bug as fixed if the script could run well. Created attachment 412 [details]
Test case for re-execution
This is my attempt at writing a script to find these bugs. It turns out to be quite tricky to cover all cases since we need to check the actual task in question reruns correctly and all subsequent tasks too. We also need to ensure sstate doesn't interfere.
(In reply to comment #5) > (In reply to comment #3) > > Jiajun, do you have the script you used to generate these results? > > Yes. I upload the script here. Could you help to check if it's OK? > I would like to add it into 1.2 M4 testing, and mark the bug as fixed if the > script could run well. I've uploaded a bit of a different version of this idea since I found it hard to catch all the different problem re-execution cases. With my version you'd do a "bitbake world" or "bitbake core-image-minimal" or whatever you'd want to test re-exection of, then run the script. It will iterate over all built stamps, re-executing tasks. I've not done much testing with it yet but it will catch the nasm-native issue we have an open bug for. Created attachment 413 [details]
Test for re-execution of tasks
Created attachment 414 [details]
Test for re-execution of tasks (correct version this time)
Created attachment 420 [details]
Test for re-execution of tasks (v4)
Attached is another updated version of this with more issues fixed. It seems to be working well now.
In order to close this bug, we need to check a script into master which has the desired functionality. At the moment, I don't like my script due to the need to patch base.bbclass and add the cleansstate2 task. So the work there isn't complete yet but I'm not sure at this point how we address that. Actually closing 2123 is not necessarily a 1.2 feature though. For 1.2 we just need to fix the bugs the script shows. Why cannot we execute cleanstate with -f option instead of adding cleanstate2 ? Created attachment 514 [details]
test-reexec
(In reply to comment #13) > Created attachment 514 [details] > test-reexec Hi Richard, I removed cleansstate2 and add a similar function in test-reexec. Could you help to review it? Patch checked in as commit 3e57b1d7494ac649c08692182aa522113926ccce. commit 3e57b1d7494ac649c08692182aa522113926ccce Author: Jiajun Xu <jiajun.xu@intel.com> Date: Fri May 18 17:18:14 2012 +0300 test-reexec: Add script to address issues when task re-execution The script is used to address build issues when tasks of different recipes are re-executed. The script goes through all available recipes and their tasks. The test results are saved in ./reexeclogs. Force build logs are saved with prefix "force". Build failure logs are saved with prefix "failed". [YOCTO #2123] (From OE-Core rev: 6258a11f22103d68d02e329c2e7fb198202cc6ef) Signed-off-by: Jiajun Xu <jiajun.xu@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Verified |