Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 16 Feb 2016 17:17:03 -0600

On 02/16/2016 04:03 PM, Michael Davidsaver wrote:
> On 02/16/2016 04:03 PM, APS Jenkins wrote:
>> Test Summary Report
>> -------------------
>> src/ioc/db/test/O.darwin-x86/callbackTest.tap           (Wstat: 0 Tests: 339 Failed: 12)
>>   Failed tests:  306-317
>> Files=72, Tests=9409,  2 wallclock secs ( 0.67 usr  0.05 sys +  0.13 cusr  0.11 csys =  0.96 CPU)
>> Result: FAIL
>> Build step 'Execute shell' marked build as failure
> 
> Ok, this one looks like a transient "error" due to system load.

Yup, that's partly why I gave it a separate -test job, so it's quick to
re-run the tests manually without having to rebuild.

>> not ok 306 - delay 11.25 seconds, callback time error |0.8907| < 0.05
>> not ok 307 - delay 11.25 seconds, callback time error |0.8906| < 0.05
>> not ok 308 - delay 11.25 seconds, callback time error |0.8904| < 0.05
>> not ok 309 - delay 11.50 seconds, callback time error |0.6407| < 0.05
>> not ok 310 - delay 11.50 seconds, callback time error |0.6406| < 0.05
>> not ok 311 - delay 11.50 seconds, callback time error |0.6404| < 0.05
>> not ok 312 - delay 11.75 seconds, callback time error |0.3907| < 0.05
>> not ok 313 - delay 11.75 seconds, callback time error |0.3906| < 0.05
>> not ok 314 - delay 11.75 seconds, callback time error |0.3904| < 0.05
>> not ok 315 - delay 12.00 seconds, callback time error |0.1407| < 0.05
>> not ok 316 - delay 12.00 seconds, callback time error |0.1406| < 0.05
>> not ok 317 - delay 12.00 seconds, callback time error |0.1404| < 0.05

0.89 seconds is an 8% over-run. We could change the tests to 10%, but I
wonder if this kind of thing is worth failing over at all?

- Andrew

-- 
There are only two hard problems in distributed systems:
  2. Exactly-once delivery
  1. Guaranteed order of messages
  2. Exactly-once delivery
 -- Mathias Verraes

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System