Feature #11113
Run more ISO testers
100%
Description
Now that we have more RAM to do so..
Subtasks
Related issues
Related to Tails - |
Resolved | 2014-02-19 | |
Blocks Tails - |
Rejected | 2016-01-27 | |
Blocked by Tails - |
Resolved | 2016-01-27 |
History
#1 Updated by intrigeri 2016-02-12 21:48:12
- blocks
Feature #11011: Design a 2nd ISO testing machine added
#2 Updated by intrigeri 2016-02-12 21:48:26
- blocked by
Feature #11010: Give lizard v2 more RAM added
#3 Updated by intrigeri 2016-02-13 01:36:10
And it’ll be time to try running all isotesters with enough RAM to store TailsToaster’s memory + the tmpfs peak usage + whatever we want to keep in disk cache e.g. the 2 ISO images + some memory for the system, that is ~23GB last time I checked (Feature #10996#note-5).
Also see Feature #10503#note-20 and follow-ups.
#4 Updated by intrigeri 2016-02-13 12:15:58
anonym: FYI I re-run a full test suite and I got 12688900 max usage in /tmp/TailsToaster
again (run du -cs /tmp/TailsToaster/ | tail -n 1 | awk '{print $1}'
in a loop, saving its output to a file, and then finding the maximum value in this file. I suspect that your measurements on Feature #10503#note-20 were including a too small subset of what takes space in there.
#5 Updated by intrigeri 2016-02-21 20:50:09
Bumped isotester 1..6 to 23GiB of RAM (24117248 KiB).
#6 Updated by intrigeri 2016-02-26 18:28:33
- related to
Feature #6729: Bump the number of CPU cores the testing VM has added
#7 Updated by intrigeri 2016-02-27 18:57:43
- blocked by
Bug #10068: Upgrade to Jenkins 2.x, using upstream packages added
#8 Updated by intrigeri 2016-02-27 20:09:22
- blocks deleted (
)Bug #10068: Upgrade to Jenkins 2.x, using upstream packages
#9 Updated by intrigeri 2016-02-27 21:35:07
- Status changed from Confirmed to In Progress
- % Done changed from 0 to 20
Set up isotester7 and 8, started benchmarking process.
#11 Updated by intrigeri 2016-02-29 02:17:23
- Blueprint changed from https://tails.boum.org/blueprint/hardware_for_automated_tests_take2/ to https://tails.boum.org/blueprint/hardware_for_automated_tests_take2/#benchmarks
#12 Updated by intrigeri 2016-02-29 13:01:07
- Status changed from In Progress to Resolved
- Assignee deleted (
intrigeri) - % Done changed from 20 to 100
- Deliverable for set to 270
To sum up my benchmark results: running 8 ISO testers + giving them more RAM + making them use tmpfs for their temporary test suite data => we’ve improved our test suite runs throughput from 3.3 to 8 runs / hour. There’s probably some more room for future improvements (Feature #6729, Feature #11175, Feature #11179) but this part of the experiment is done, and is a great success.