On 16/07/17 16:47, Ben Caradoc-Davies wrote:
On 16/07/17 15:58, Ben Caradoc-Davies wrote:
To avoid timeouts, I changed Travis CI to use -T1C, expecting about
four threads. Instead, Maven reports:
"Using the MultiThreadedBuilder implementation with a thread count of 32"
Yes, not two, but 32 threads, despite the documented 2 cores:
https://docs.travis-ci.com/user/ci-environment/
This might cause worker VM memory exhaustion, as each thread can get
its own test JVM, and workers are documented to have only 4 GB of memory.
I will see if -T2 yields better results.
Setting -T2 fixed a failing geoserver build. I am trying -T4 to see if
it is faster.
Whatever is used in the containerised environment to restrict resource
usage, it is not exposed via the usual operating system interface and so
-T1C causes 32 threads and risks breaching memory limits (if enough
modules are built and tested concurrently; this is a race condition).
The documented limit is 2 CPUs and 4 GB of memory.
I have changed all GeoTools and GeoServer Travis CI builds to use -T4.
Hopefully that is in a sweet spot that is fast enough to not time out
yet not so memory-greedy to get killed.
Kind regards,
--
Ben Caradoc-Davies <[email protected]>
Director
Transient Software Limited <http://transient.nz/>
New Zealand
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Geoserver-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geoserver-devel
