0

To follow up on Chuck's testing I had originally observed the threaderciser
issue on macOS 10.14 in my first round of testing a few days ago when I ran
4 sets of tests and all had only 1 failure - c_threaderciser.
In retesting this morning on the same platform with the same build with no
reboots or anything in between all tests pass including c_threaderciser, so
I can confirm the intermittent nature of the failure as noted in
PROTON-2259 [1].

In testing the new examples, raw_echo and raw_connect I saw that raw_echo
worked ok but raw_connect consistently reported:

i7mbp:examples rkieley$ ./raw_connect
Segmentation fault: 11
i7mbp:examples rkieley$

While the regular broker/receive/send, broker/simple_recv/simple_send and
others all operated as expected.

I think perhaps an rc2 is warranted, however haven't had a chance to
identify earlier and help address the following:
- the raw connection work is a big feature addition and may just require a
small update
- the threaderciser work via PROTON-2225 had ' Set -DTHREADERCISER=ON by
default, to prevent regressions' [1] so it seems fair to expect that it
would help reduce problems.

As such I don't feel strongly enough to -1


Roddie
---
[1] - https://issues.apache.org/jira/browse/PROTON-2259
[2] - https://issues.apache.org/jira/browse/PROTON-2225

On Fri, Aug 21, 2020 at 11:40 AM Chuck Rolke <[email protected]> wrote:

> Change to +1
>
> Chalking up test failures to be a "test problem" and not a "proton
> problem".
>
> Local test setup first system, the one where the test fails:
>
>  * ThinkPad P50 laptop
>  * Linux unused.localdomain 5.7.9-100.fc31.x86_64 #1 SMP Fri Jul 17
> 17:18:38 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
>
>  * Testbed ran overnight - no issues
>  * c_threaderciser standalone C executable from command prompt:
>      runs hundreds of times - no issues
>  * c_threaderciser under ctest with valgrind-3.16.0, memcheck, leak check:
>      passes sometimes but usually times out
>
> Local test setup second system:
>
>  * Dell Inspiron desktop
>  * Linux taj.localdomain 5.7.11-100.fc31.x86_64 #1 SMP Wed Jul 29 18:17:53
> UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
>
>  * c_threaderciser under ctest with valgrind-3.16.0 passes every time
>      ctest reporting elapsed time around 3 seconds,
>
> Problem system 'unused' has a significant difference from other systems
> with an
> NVIDIA Quadro M1000M GPU running NVIDIA Driver 450.57.
>
>
> ----- Original Message -----
> > From: "Robbie Gemmell" <[email protected]>
> > To: [email protected]
> > Sent: Thursday, August 20, 2020 5:57:18 PM
> > Subject: Re: [VOTE] Release Apache Qpid Proton 0.32.0
> >
> > On Thu, 20 Aug 2020 at 20:25, Andrew Stitcher <[email protected]>
> wrote:
> > >
> > > On Thu, 2020-08-20 at 14:50 -0400, Chuck Rolke wrote:
> > > > -1
> > > >
> > > > * checksum matches
> > > > * built debug build
> > > > * threaderciser self test fails hard with timeout
> > > >
> > > > Issues with the test:
> > > >
> > > > - test does not honor ctest --timeout switch. Test always uses 120
> > > > seconds.
> > > > - test times out with Debug, RelWithDebInfo, and Release builds
> > > >
> > > > Investigating a little:
> > > >   - Sometimes takes 30 seconds just to *start* 8 threads
> > > >   - Test times out with millisleep commented out.
> > > >   -- sometimes with no user threads returning from join
> > > >   -- sometimes user threads join but proactor threads do not
> > >
> > > Why would this be a reason to block the release if the bug that reports
> > > this behaviour [1] was not a blocker 3 months ago?
> > >
> > > Andrew
> > >
> > > [1] https://issues.apache.org/jira/browse/PROTON-2230
> > >
> >
> > To be clear: a -1 on a release vote is not a block, they can't be
> > vetoed. If there are sufficient binding +1's and a net positive vote,
> > it is the release managers choice how to proceed based on issues
> > presented. So please keep testing and voting folks.
> >
> > In this case I'll essentially rely on those who would be fixing any
> > such issues to creating an argument/option not to proceed. If there
> > are no imminent changes/fix, and there doesn't seem to have been a new
> > change in behaviour from months ago, then so long as the vote tally
> > allows the release is currently likely to proceed.
> >
> > Chuck says it fails every time for him, which is interesting as it
> > hasn't done so elsewhere that I'm aware. Especially in a multitude of
> > builds/envs run over the last week that I have been looking to
> > progress with the release. That might make Chucks env useful for
> > nailing down any changes here.
> >
> > > >
> > > >
> > > > ----- Original Message -----
> > > > > From: "Robbie Gemmell" <[email protected]>
> > > > > To: [email protected]
> > > > > Sent: Thursday, August 20, 2020 6:29:22 AM
> > > > > Subject: [VOTE] Release Apache Qpid Proton 0.32.0
> > > > >
> > > > > Hi folks,
> > > > >
> > > > > I have put together a spin for a Qpid Proton 0.32.0 release, please
> > > > > give it a test out and vote accordingly.
> > > > >
> > > > > The files can be grabbed from:
> > > > > https://dist.apache.org/repos/dist/dev/qpid/proton/0.32.0-rc1/
> > > > >
> > > > > The JIRAs assigned are:
> > > > >
> https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12313720&version=12348172
> > > > >
> > > > > It is tagged as 0.32.0-rc1.
> > > > >
> > > > > Regards,
> > > > > Robbie
> > > > >
> > > > > -----------------------------------------------------------------
> > > > > ----
> > > > > To unsubscribe, e-mail: [email protected]
> > > > > For additional commands, e-mail: [email protected]
> > > > >
> > > > >
> > > >
> > > > ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: [email protected]
> > > > For additional commands, e-mail: [email protected]
> > > >
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [email protected]
> > > For additional commands, e-mail: [email protected]
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [email protected]
> > For additional commands, e-mail: [email protected]
> >
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>

Reply via email to