Bug 178019 – Layout test http/tests/cache/disk-cache/disk-cache-validation-no-body.html is timing out.

• Status: RESOLVED
• Resolution: CONFIGURATION CHANGED
• Component: New Bugs

[Matt Lewis]

http/tests/cache/disk-cache/disk-cache-validation-no-body.html is timing out on El Capitan Release according to the flakiness dashboard.

https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=http%2Ftests%2Fcache%2Fdisk-cache%2Fdisk-cache-validation-no-body.html

https://build.webkit.org/results/Apple%20El%20Capitan%20Release%20WK2%20(Tests)/r222977%20(5032)/results.html
https://build.webkit.org/builders/Apple%20El%20Capitan%20Release%20WK2%20(Tests)/builds/5032

This test has already been marked as Flaky on macOS Debug WK2 with related bug. 

The first timeout on El Capitan Release was with:
https://build.webkit.org/results/Apple%20El%20Capitan%20Release%20WK2%20(Tests)/r222866%20(4962)/results.html
https://build.webkit.org/builders/Apple%20El%20Capitan%20Release%20WK2%20(Tests)/builds/4962

[Alexey Proskuryakov]

Is this a regression? I thought that all the tests were supposed to pass in release builds now.

[Matt Lewis]

It looks like after the test was skipped in https://trac.webkit.org/changeset/222994/webkit

These tests started to time out as well:
http/tests/cache/disk-cache/disk-cache-vary-no-body.html
http/tests/cache/disk-cache/disk-cache-vary.html

I am testing locally to see if the timeout is still occurring and trying to bisect a possible regression point.

[Antti Koivisto]

Looking at that dashboard numbers, the running time of the test on El Capitan bot has been steadily increasing over the last few weeks until it crossed the 30s threshold.

[Antti Koivisto]

Maybe the bot is choking on something?

[Matt Lewis]

The two test I mentioned before also seem to have go through that process of steadily growing in time.

[Alexey Proskuryakov]

Good catch. There is a known kernel bug (fixed in High Sierra) making processes on WebKit test bots launch slower over time, which has particularly noticeable effect on http tests.

These bots have uptime of 21 days, which used to be OK. Perhaps something changed that makes the problem manifest faster now.

I suggest unskipping the test, rebooting the bots, and watching this test closely for the next few weeks.

[Matt Lewis]

Unskipped in https://trac.webkit.org/changeset/223071/webkit

[Ryan Haddad]

After the unskipping and rebooting, it looks like things are good again on the El Capitan Release bots. Moving this to resolved / config changed.

[Radar WebKit Bug Importer]

<rdar://problem/34899513>