Re: [VOTE] Release Qpid Proton 0.15.0

2016-10-11 Thread Timothy Bish
+1 * Validated signatures and checksums * Built from source using CMake and Maven (proton-j) and ran tests * Tested ActiveMQ broker using the staged libraries On 10/10/2016 01:20 PM, Robbie Gemmell wrote: Hi folks, I have put together a spin for a 0.15.0 Qpid Proton release as proposed,

Re: [VOTE] Release Qpid Dispatch Router 0.7.0 (RC1)

2016-10-11 Thread Ganesh Murthy
-1 kgiusti ran a coverity scan on RC1 which found 7 memory leaks. These need to be fixed. I have entered a JIRA with the details - https://issues.apache.org/jira/browse/DISPATCH-526 Thanks. - Original Message - > From: "Ted Ross" > To: users@qpid.apache.org >

Re: [Dispatch router 0.6.2] Performance regression of the producers throughput

2016-10-11 Thread Adel Boutros
Hello Ted, As there are not a lot of fixes in 0.6.2, would you like me to identify the culprit fix or do you prefer I test directly on 0.7.0? For your information, we are currently compiling Proton 0.14.0 so we can use the 0.7.0 but we still need some time before our code has been adapted.

Re: qpid.work_dir property is not honoring on broker startup

2016-10-11 Thread Robbie Gemmell
On 11 October 2016 at 11:01, Rob Godfrey wrote: > On 11 October 2016 at 10:45, Robbie Gemmell wrote: >> From memory of how it used to [be intended to] work long ago, >> qpid.work_dir was a config property that could be set to influence the >>

Re: [Proton-c][Solaris][SunStudio-2.1] Compiling Qpid Proton-c and C++ bindings on Solaris

2016-10-11 Thread Adel Boutros
Hello again, I have added one final fix listed below. For your information, I have replayed the patches at head of master on Linux (GCC 4.9.1) and Solaris (SunStudio 12.1). The compilation is successful and all unit tests are green on both configuration for Proton-c and the C++ bindings.

Re: qpid.work_dir property is not honoring on broker startup

2016-10-11 Thread Rob Godfrey
On 11 October 2016 at 10:45, Robbie Gemmell wrote: > From memory of how it used to [be intended to] work long ago, > qpid.work_dir was a config property that could be set to influence the > work dir actually used and was referenced in the config json etc > accordingly,

Re: qpid.work_dir property is not honoring on broker startup

2016-10-11 Thread Robbie Gemmell
>From memory of how it used to [be intended to] work long ago, qpid.work_dir was a config property that could be set to influence the work dir actually used and was referenced in the config json etc accordingly, but if its value was not set explicitly then the broker would fall back to the value

Re: qpid.work_dir property is not honoring on broker startup

2016-10-11 Thread Rob Godfrey
Looking at the code, in the startup phase the Broker seems to use only QPID_WORK as provided in a system property, or qpid.work_dir provided on the command line with the -prop qpid.work_dir=foo form. It seems that it does not try to look for qpid.work_dir in the system properties (or, implicitly,

Re: qpid.work_dir property is not honoring on broker startup

2016-10-11 Thread Lorenz Quack
Hi Ram, Notice that in your example both QPID_WORK and qpid.work_dir are specified. It seems that currently QPID_WORK take precedence. I guess if the environment variable and system property QPID_WORK are not set then the broker picks up the qpid.work_dir property, right? Kind regards, Lorenz