You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by Alan Conway <ac...@redhat.com> on 2011/04/19 20:31:08 UTC
Proposed fixes for 0.10 (was Re: What is our strategy for "errata"
and/or "hot fixes" ?)
On 04/19/2011 12:26 PM, Rajith Attapattu wrote:
> Based on the comments so far it seems that delaying the 0.10 release
> is probably a better idea.
> I agree with Carl and Steve that erratas can be time consuming and may
> actually be detrimental to our overall goals.
>
> For now lets hold on to the release a bit and see what needs to be
> done to fix QPID-3214& QPID-3216.
> Based on that we can make a call to just release note it and fix it in
> 0.12 or whether we can safely fix them for 0.10
>
If we are holding up 0.10 release I'd like to propose these two fixes:
QPID-3215 cached exchange reference can cause cluster inconsistencies if
exchange is deleted/recreated
QPID-3208 Exchanges make best effort to route messages if there is an error.
They are not regressions but both can cause cluster brokers to shut down (and
have for at least one user.) The fixes are low risk and already on the trunk.
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org
Re: Proposed fixes for 0.10 (was Re: What is our strategy for "errata"
and/or "hot fixes" ?)
Posted by Justin Ross <jr...@redhat.com>.
On Wed, 20 Apr 2011, Alan Conway wrote:
> On 04/19/2011 02:31 PM, Alan Conway wrote:
>> If we are holding up 0.10 release I'd like to propose these two fixes:
>>
>> QPID-3215 cached exchange reference can cause cluster inconsistencies if
>> exchange is deleted/recreated
>> QPID-3208 Exchanges make best effort to route messages if there is an
>> error.
>>
>> They are not regressions but both can cause cluster brokers to shut down
>> (and
>> have for at least one user.) The fixes are low risk and already on the
>> trunk.
>>
>
> I'd like to add this trivial fix:
>
> QPID-3217 Exchanges with IVE option cause cluster inconsistencies in updatees
>
> All 3 are on trunk and have had code review.
Hi, Alan. It's looking more and more like we will not extend 0.10 and not
produce a new RC, so these will probably need to wait for 0.12.
Justin
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org
Re: Proposed fixes for 0.10 (was Re: What is our strategy for "errata"
and/or "hot fixes" ?)
Posted by Jonathan Robie <jo...@redhat.com>.
I have one change for this list:
QPID-3185, see the second comment
qpid-config has two fixes on trunk that are not in the 0.10 release
candidate:
1. Fixing the text for --broker-addr, which had the text for the wrong
option
2. Fixing the name of --force-if-used, which had been mistakenly renamed
--force-if-not-used
Jonathan
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org
Re: Proposed fixes for 0.10 (was Re: What is our strategy for "errata"
and/or "hot fixes" ?)
Posted by Alan Conway <ac...@redhat.com>.
On 04/20/2011 09:43 AM, Alan Conway wrote:
> On 04/19/2011 02:31 PM, Alan Conway wrote:
>> On 04/19/2011 12:26 PM, Rajith Attapattu wrote:
>>> Based on the comments so far it seems that delaying the 0.10 release
>>> is probably a better idea.
>>> I agree with Carl and Steve that erratas can be time consuming and may
>>> actually be detrimental to our overall goals.
>>>
>>> For now lets hold on to the release a bit and see what needs to be
>>> done to fix QPID-3214& QPID-3216.
>>> Based on that we can make a call to just release note it and fix it in
>>> 0.12 or whether we can safely fix them for 0.10
>>>
>>
>> If we are holding up 0.10 release I'd like to propose these two fixes:
>>
>> QPID-3215 cached exchange reference can cause cluster inconsistencies if
>> exchange is deleted/recreated
>> QPID-3208 Exchanges make best effort to route messages if there is an error.
>>
>> They are not regressions but both can cause cluster brokers to shut down (and
>> have for at least one user.) The fixes are low risk and already on the trunk.
>>
>
> I'd like to add this trivial fix:
>
> QPID-3217 Exchanges with IVE option cause cluster inconsistencies in updatees
>
> All 3 are on trunk and have had code review.
I'd like to add 2 more - this is the final list:
QPID-3208 Exchanges make best effort to route messages if there is an error.
QPID-3215 cached exchange reference can cause cluster inconsistencies if
exchange is deleted/recreated (2)
QPID-3217 Exchanges with IVE option cause cluster inconsistencies in updatees
QPID-3221 Fix bug in Cluster::timerDrop - calling wrong function
QPID-3202: Clustered broker shuts down with "unknown connection" error
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org
Re: Proposed fixes for 0.10 (was Re: What is our strategy for "errata"
and/or "hot fixes" ?)
Posted by Alan Conway <ac...@redhat.com>.
On 04/19/2011 02:31 PM, Alan Conway wrote:
> On 04/19/2011 12:26 PM, Rajith Attapattu wrote:
>> Based on the comments so far it seems that delaying the 0.10 release
>> is probably a better idea.
>> I agree with Carl and Steve that erratas can be time consuming and may
>> actually be detrimental to our overall goals.
>>
>> For now lets hold on to the release a bit and see what needs to be
>> done to fix QPID-3214& QPID-3216.
>> Based on that we can make a call to just release note it and fix it in
>> 0.12 or whether we can safely fix them for 0.10
>>
>
> If we are holding up 0.10 release I'd like to propose these two fixes:
>
> QPID-3215 cached exchange reference can cause cluster inconsistencies if
> exchange is deleted/recreated
> QPID-3208 Exchanges make best effort to route messages if there is an error.
>
> They are not regressions but both can cause cluster brokers to shut down (and
> have for at least one user.) The fixes are low risk and already on the trunk.
>
I'd like to add this trivial fix:
QPID-3217 Exchanges with IVE option cause cluster inconsistencies in updatees
All 3 are on trunk and have had code review.
---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project: http://qpid.apache.org
Use/Interact: mailto:dev-subscribe@qpid.apache.org