Failures on sandbox under buildbot

[wd15]

The parallel failure is nor reproducible so better off leaving that for the time being.

http://build.cmi.kent.edu:8010/builders/Ubuntu_x86_64~trunk~full/builds/15/steps/trial_2/logs/stdio

The inline seg fault can't be reproduced at the command line so not going to bother with that either.

http://build.cmi.kent.edu:8010/builders/Ubuntu_x86_64~trunk~full/builds/15/steps/trial_4/logs/stdio

The timeout in the pyamg tests is down to the increase in time of the Cahn-Hilliard mesh2D.py test. That test takes 2187.809 on bunter. It needs to be below 1200 so removing the preconditioner for that one test should be adequate as the scipy solver alone must be under 1200s.

http://build.cmi.kent.edu:8010/builders/Ubuntu_x86_64~trunk~full/builds/15/steps/trial_6/logs/stdio

Imported from trac ticket #427, created by wd15 on 01-30-2012 at 17:06, last modified: 01-31-2012 at 14:37

[wd15]

r5138 should deal with the slow solution of mesh2D.py with the pyamg solvers.

Trac comment by wd15 on 01-31-2012 at 12:43

[guyer]

The first two failures have disappeared and r5138 resolved the PyAMG issues

Trac comment by guyer on 01-31-2012 at 14:37