Details
-
Bug
-
Resolution: Done
-
Major
-
master
-
ubuntu12-cv-02 (http://cv.jenkins.couchbase.com/computer/ubuntu12-cv-02/)
2016-05-25 07:23:28 (machine time zone)
-
Untriaged
-
Ubuntu 64-bit
-
Unknown
Description
I had a commit-validation job timeout today - http://cv.jenkins.couchbase.com/job/ep-engine-threadsanitizer-master/760/consoleFull#2002939979c5751633-7aa5-473a-91e8-426e0afa5a7b Looking at the build log, I saw that the test which failed timed out after waiting for a file to exist (after 5 seconds).
This is a generous timeout - the test takes 53ms on my local Linux server (24core Xeon, spinning disks):
[ RUN ] FullAndValueEviction/EPStoreEvictionTest.GetKeyStatsEjected/value_only
|
[ OK ] FullAndValueEviction/EPStoreEvictionTest.GetKeyStatsEjected/value_only (53 ms)
|
Whereas on the build slave:
[ RUN ] FullAndValueEviction/EPStoreEvictionTest.GetKeyStatsEjected/value_only
|
/home/couchbase/jenkins/workspace/ep-engine-threadsanitizer-master/ep-engine/tests/module_tests/evp_store_test.cc:186: Failure
|
Value of: result
|
Actual: -1
|
Expected: 1
|
Failed to flush the one item we have stored.
|
[ FAILED ] FullAndValueEviction/EPStoreEvictionTest.GetKeyStatsEjected/value_only, where GetParam() = "value_only" (6375 ms)
|
(Note this test is running under threadSanitizer, but that's true in both environments.)
I don't have an issue with tests running a bit slower on the commit-validation machines, but over two orders of magnitude seems excessive.
Could someone from build / IT take a look at this?