I tried cleaning as well. It turns out that the directory I needed to
remove was `~/.m2`. I guess the package cache was corrupted. Transient
failure?
On Tue, Jan 22, 2019 at 5:49 PM Galen O'Sullivan
wrote:
> I'm getting the following failure building Geode on the latest develop.
> I've tried `rm
I saw the same problem a few weeks ago. I ended up deleting the directories
in my .m2 repository and rebuilding. That seemed to fix it.
The cause seems to have something to do with that log4j core tests jar, but
I'm not sure why our build would be looking for the corresponding sources
jar. If anyo
You might also try ./gradlew build --refresh-dependencies if this happens
again.
On Wed, Jan 23, 2019 at 9:38 AM Kirk Lund wrote:
> I saw the same problem a few weeks ago. I ended up deleting the directories
> in my .m2 repository and rebuilding. That seemed to fix it.
>
> The cause seems to hav
I hit a dunit hang in one of my precheckin runs.
The only test mentioned in callstacks/dunit-hangs.txt is
ClusterConfigLocatorRestartDUnitTest.
I see some Pooled Message Processor threads that might be hung waiting for
the same
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject
Looks like there's a Test worker thread that keeps running and changing.
Maybe the dunit test is pushing the dunit job over its time limit:
"Test worker" #27 prio=5 os_prio=0 cpu=1332.79ms elapsed=4871.91s
tid=0x7f67c4a78800 nid=0x1c runnable [0x7f675fbf3000]
java.lang.Thread.State: RU
Thanks Kirk - this is GEODE-5676. I'm glad the callstacks are finally
working again, that should help us track this down! I'll update that ticket
with anything I can figure out from this failure.
-Dan
On Wed, Jan 23, 2019 at 10:05 AM Kirk Lund wrote:
> Looks like there's a Test worker thread th
Dan, 5676 is for the other test in the same class. I'm working on that
other test as part of fixing GEODE-6309.
On 1/23/19 10:43 AM, Dan Smith wrote:
Thanks Kirk - this is GEODE-5676. I'm glad the callstacks are finally
working again, that should help us track this down! I'll update that ticke
There is a ticket open for a hang in the other test method in that same
class.
On 1/23/19 10:04 AM, Kirk Lund wrote:
Looks like there's a Test worker thread that keeps running and changing.
Maybe the dunit test is pushing the dunit job over its time limit:
"Test worker" #27 prio=5 os_prio=0 cp
Thanks, Jared. I'll try that next time.
On Wed, Jan 23, 2019 at 9:53 AM Jared Stewart
wrote:
> You might also try ./gradlew build --refresh-dependencies if this happens
> again.
>
> On Wed, Jan 23, 2019 at 9:38 AM Kirk Lund wrote:
>
> > I saw the same problem a few weeks ago. I ended up deletin
Ryan, thank you so much for the great explanation of your proposal!
This seems very sound and you and David got me convinced that it's the
right thing to change the default. To me the question now is one of timing.
Is this something we can change in a minor release or do we have to wait
for Geode
10 matches
Mail list logo
