Re: Failure to push message seen in a dunit test

Yep, the test method was testReplaceOfflineMemberAndRestart_WithMultipleDiskStores. I can probably reproduce with log files if that helps. On Wed, Feb 19, 2020 at 10:04 AM Bruce Schuchardt wrote: > Kirk, do you recall which of the tests in that class hit this problem? It > looks pretty serious.

Re: Redis PubSubTest started failing

Not at this time since I do believe the issue is just a flaky test and not a flaky implementation. I've merged in the change to ignore this test for now. --Jens On Wed, Feb 19, 2020 at 9:46 AM Udo Kohlmeyer wrote: > Does that mean we will revert from develop pls? > > On 2/19/20 9:33 AM, Jens D

Odg: Question about Hash Indexes

Thanks for clarification. Šalje: Jason Huynh Poslano: 19. veljače 2020. 18:05 Prima: geode Predmet: Re: Question about Hash Indexes Hi Mario, >From my understanding: 1.) The Hash Index was implemented unlike a traditional hash index, instead it is more of a memo

Re: Failure to push message seen in a dunit test

Kirk, do you recall which of the tests in that class hit this problem? It looks pretty serious. On 2/19/20, 9:24 AM, "Kirk Lund" wrote: While running PersistentColocatedPartitionedRegionDistributedTest a thousand times to verify that I've fixed a flaky issue in the test, it hit an

Re: Redis PubSubTest started failing

Does that mean we will revert from develop pls? On 2/19/20 9:33 AM, Jens Deppe wrote: Me too :) I've had it pass several StressNewTest runs but yet, it seems to still fail at times. On Wed, Feb 19, 2020 at 9:29 AM Kirk Lund wrote: PubSubTest now fails for me in every PR I submit (3 yesterda

Re: Redis PubSubTest started failing

I've created a PR to ignore this test for now (GEODE-7798 is tracking the problem). On Wed, Feb 19, 2020 at 9:33 AM Jens Deppe wrote: > Me too :) > > I've had it pass several StressNewTest runs but yet, it seems to still > fail at times. > > On Wed, Feb 19, 2020 at 9:29 AM Kirk Lund wrote: > >>

Re: Redis PubSubTest started failing

Me too :) I've had it pass several StressNewTest runs but yet, it seems to still fail at times. On Wed, Feb 19, 2020 at 9:29 AM Kirk Lund wrote: > PubSubTest now fails for me in every PR I submit (3 yesterday). I'm really > curious how this change made it past PR and into develop. > > On Wed, F

Re: Redis PubSubTest started failing

PubSubTest now fails for me in every PR I submit (3 yesterday). I'm really curious how this change made it past PR and into develop. On Wed, Feb 19, 2020 at 9:18 AM Kirk Lund wrote: > PubSubTest is an integration test. Confusion from test names is why I > prefer to include *IntegrationTest or *D

Failure to push message seen in a dunit test

While running PersistentColocatedPartitionedRegionDistributedTest a thousand times to verify that I've fixed a flaky issue in the test, it hit an interesting failure trying to send a RequestImageMessage. This generated a stack trace which caused the test to fail grep for suspect strings. I can easi

Re: Redis PubSubTest started failing

PubSubTest is an integration test. Confusion from test names is why I prefer to include *IntegrationTest or *DistributionTest in the name like the wiki prescribes. I recommend naming this test PubSubIntegrationTest. On Wed, Feb 19, 2020 at 9:06 AM Owen Nichols wrote: > StressNew is cross-cutting

Re: Redis PubSubTest started failing

@Owen.. off topic.. but i think it is worth the discussion On 2/19/20 9:00 AM, Owen Nichols wrote: Perhaps also worth considering: can we get WindowsStressNew added to the PR checks? On Wed, Feb 19, 2020 at 8:50 AM Udo Kohlmeyer wrote: Is this something that can be fixed in a short time (2hr

Re: Redis PubSubTest started failing

StressNew is cross-cutting. Unit tests are historically the very least likely to exhibit windows-specific modes of failure. On Wed, Feb 19, 2020 at 9:02 AM Robert Houghton wrote: > We don't even have windows unit tests for PRs. Walk before we run... > > On Wed, Feb 19, 2020, 09:00 Owen Nichols

Re: Question about Hash Indexes

Hi Mario, >From my understanding: 1.) The Hash Index was implemented unlike a traditional hash index, instead it is more of a memory saving index that uses hashing to not store keys. When the backing data structure needs to expand, it needs to rehash a lot of data and this can be detrimental at ru

Re: Redis PubSubTest started failing

We don't even have windows unit tests for PRs. Walk before we run... On Wed, Feb 19, 2020, 09:00 Owen Nichols wrote: > Perhaps also worth considering: can we get WindowsStressNew added to the PR > checks? > > On Wed, Feb 19, 2020 at 8:50 AM Udo Kohlmeyer wrote: > > > Is this something that can

Re: Redis PubSubTest started failing

Perhaps also worth considering: can we get WindowsStressNew added to the PR checks? On Wed, Feb 19, 2020 at 8:50 AM Udo Kohlmeyer wrote: > Is this something that can be fixed in a short time (2hrs)? If not, can > be revert and get back to a clean pipeline? > > --Udo > > On 2/19/20 8:23 AM, Jens

Re: Redis PubSubTest started failing

Is this something that can be fixed in a short time (2hrs)? If not, can be revert and get back to a clean pipeline? --Udo On 2/19/20 8:23 AM, Jens Deppe wrote: Thanks Kirk, We're working on fixing this. --Jens On Tue, Feb 18, 2020 at 3:23 PM Kirk Lund wrote: I just started seeing the Red

Re: Redis PubSubTest started failing

Thanks Kirk, We're working on fixing this. --Jens On Tue, Feb 18, 2020 at 3:23 PM Kirk Lund wrote: > I just started seeing the Redis PubSubTest fail in IntegrationTest after > rebasing on develop this afternoon. Looks like I have Jens' latest commit > for this test: > > > > > > > > *commit 1be

Question about Hash Indexes

Hi geode dev, could you help us with some questions regarding Hash Indexes: 1. Why are Hash Indexes deprecated? 2. Are there any plans in the future for there removal? Thanks, Mario