You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@storm.apache.org by "P. Taylor Goetz" <pt...@gmail.com> on 2015/06/08 23:18:52 UTC
Blog Post for Storm 0.10.0 Release
I’m working on a blog post for the 0.10.0 release and plan on highlighting the main new features. There are a lot and I want to make sure I mention them. Below is what I have off the top of my head (in no particular order aside from the fist):
* Security (the big one)
* Support for future rolling upgrades
* Kafka integration improvements
* Hive integration
* MS Azure Eventhubs Integration
* Redis integration
* JDBC integration
* Partial key grouping
* Dependency conflict reduction
* Topology deployment and testing tool (flux)
* New logging framework
If you think I missed anything let me know. And if you think of any important points for any of the above that you’d like to see included, let me know.
I’d also like to include a list of changes that are not backward compatible. I’m not sure we need it for the announcement, but we should definitely add it to the documentation. We don’t currently track that information in JIRA (we should consider doing so), so if anyone can help with compiling a list, I’d appreciate it.
I will try to send out a draft for review tomorrow.
And most importantly, thanks and congratulations to everyone who contributed to this release, it’s a big one!
-Taylor
Re: Blog Post for Storm 0.10.0 Release
Posted by Bobby Evans <ev...@yahoo-inc.com.INVALID>.
It looks good to me. There are a few spelling mistakes, that TextMate pointed out to me, but beyond that it looks fine.
- Bobby
On Friday, June 12, 2015 4:26 PM, P. Taylor Goetz <pt...@gmail.com> wrote:
Here’s the latest working draft of the blog post, which I hope to post on Monday following the release.
(note that I haven’t finished the shaded dependency table)
Any feedback is welcome.
-Taylor
On Jun 8, 2015, at 5:18 PM, P. Taylor Goetz <pt...@gmail.com> wrote:
> I’m working on a blog post for the 0.10.0 release and plan on highlighting the main new features. There are a lot and I want to make sure I mention them. Below is what I have off the top of my head (in no particular order aside from the fist):
>
> * Security (the big one)
> * Support for future rolling upgrades
> * Kafka integration improvements
> * Hive integration
> * MS Azure Eventhubs Integration
> * Redis integration
> * JDBC integration
> * Partial key grouping
> * Dependency conflict reduction
> * Topology deployment and testing tool (flux)
> * New logging framework
>
> If you think I missed anything let me know. And if you think of any important points for any of the above that you’d like to see included, let me know.
>
> I’d also like to include a list of changes that are not backward compatible. I’m not sure we need it for the announcement, but we should definitely add it to the documentation. We don’t currently track that information in JIRA (we should consider doing so), so if anyone can help with compiling a list, I’d appreciate it.
>
> I will try to send out a draft for review tomorrow.
>
> And most importantly, thanks and congratulations to everyone who contributed to this release, it’s a big one!
>
> -Taylor
>
Re: Blog Post for Storm 0.10.0 Release
Posted by "P. Taylor Goetz" <pt...@gmail.com>.
Here’s the latest working draft of the blog post, which I hope to post on Monday following the release.
(note that I haven’t finished the shaded dependency table)
Any feedback is welcome.
-Taylor
Re: Blog Post for Storm 0.10.0 Release
Posted by Bobby Evans <ev...@yahoo-inc.com.INVALID>.
Ya the reason I have been thinking about it is because we have been trough it with our customers for security already, and we have are preparing for upmerging to 0.10.0 and we need to tell our customers what to do.
- Bobby
On Monday, June 8, 2015 5:15 PM, P. Taylor Goetz <pt...@gmail.com> wrote:
I'm thinking also about external components. I'm not in front of source right now, but I think we changed the kafka config in a non backward compatible way.
We're probably okay with a statement about recompiling. And, yes, older components such as storm-kafka should continue to work without repackaging. I just want to be thorough.
So what you mentioned is exactly what I'm looking for (and something I'd forgotten about).
-Taylor
> On Jun 8, 2015, at 6:00 PM, Bobby Evans <ev...@yahoo-inc.com.INVALID> wrote:
>
> What do you mean by not backwards compatible? From the API perspective it should be binary compatible, and should be able to run old topologies without a recompile (assuming no dependencies from storm were used, and I believe at least one config was removed, but you would have to check). When recompiling a new NotAuthorizedException can be thrown by some APIs that will need to be handled. From an end user perspective the big difference is in the dependencies, especially around shading of dependencies. This means that end users will want to recompile their topologies using at least 0.10.0, so that maven packages their dependencies in the uber jar correctly.
>
> The REST API should still be backwards compatible.
>
> - Bobby
>
>
>
> On Monday, June 8, 2015 4:19 PM, P. Taylor Goetz <pt...@gmail.com> wrote:
>
>
> I’m working on a blog post for the 0.10.0 release and plan on highlighting the main new features. There are a lot and I want to make sure I mention them. Below is what I have off the top of my head (in no particular order aside from the fist):
>
> * Security (the big one)
> * Support for future rolling upgrades
> * Kafka integration improvements
> * Hive integration
> * MS Azure Eventhubs Integration
> * Redis integration
> * JDBC integration
> * Partial key grouping
> * Dependency conflict reduction
> * Topology deployment and testing tool (flux)
> * New logging framework
>
> If you think I missed anything let me know. And if you think of any important points for any of the above that you’d like to see included, let me know.
>
> I’d also like to include a list of changes that are not backward compatible. I’m not sure we need it for the announcement, but we should definitely add it to the documentation. We don’t currently track that information in JIRA (we should consider doing so), so if anyone can help with compiling a list, I’d appreciate it.
>
> I will try to send out a draft for review tomorrow.
>
> And most importantly, thanks and congratulations to everyone who contributed to this release, it’s a big one!
>
> -Taylor
>
>
Re: Blog Post for Storm 0.10.0 Release
Posted by "P. Taylor Goetz" <pt...@gmail.com>.
I'm thinking also about external components. I'm not in front of source right now, but I think we changed the kafka config in a non backward compatible way.
We're probably okay with a statement about recompiling. And, yes, older components such as storm-kafka should continue to work without repackaging. I just want to be thorough.
So what you mentioned is exactly what I'm looking for (and something I'd forgotten about).
-Taylor
> On Jun 8, 2015, at 6:00 PM, Bobby Evans <ev...@yahoo-inc.com.INVALID> wrote:
>
> What do you mean by not backwards compatible? From the API perspective it should be binary compatible, and should be able to run old topologies without a recompile (assuming no dependencies from storm were used, and I believe at least one config was removed, but you would have to check). When recompiling a new NotAuthorizedException can be thrown by some APIs that will need to be handled. From an end user perspective the big difference is in the dependencies, especially around shading of dependencies. This means that end users will want to recompile their topologies using at least 0.10.0, so that maven packages their dependencies in the uber jar correctly.
>
> The REST API should still be backwards compatible.
>
> - Bobby
>
>
>
> On Monday, June 8, 2015 4:19 PM, P. Taylor Goetz <pt...@gmail.com> wrote:
>
>
> I’m working on a blog post for the 0.10.0 release and plan on highlighting the main new features. There are a lot and I want to make sure I mention them. Below is what I have off the top of my head (in no particular order aside from the fist):
>
> * Security (the big one)
> * Support for future rolling upgrades
> * Kafka integration improvements
> * Hive integration
> * MS Azure Eventhubs Integration
> * Redis integration
> * JDBC integration
> * Partial key grouping
> * Dependency conflict reduction
> * Topology deployment and testing tool (flux)
> * New logging framework
>
> If you think I missed anything let me know. And if you think of any important points for any of the above that you’d like to see included, let me know.
>
> I’d also like to include a list of changes that are not backward compatible. I’m not sure we need it for the announcement, but we should definitely add it to the documentation. We don’t currently track that information in JIRA (we should consider doing so), so if anyone can help with compiling a list, I’d appreciate it.
>
> I will try to send out a draft for review tomorrow.
>
> And most importantly, thanks and congratulations to everyone who contributed to this release, it’s a big one!
>
> -Taylor
>
>
Re: Blog Post for Storm 0.10.0 Release
Posted by Bobby Evans <ev...@yahoo-inc.com.INVALID>.
What do you mean by not backwards compatible? From the API perspective it should be binary compatible, and should be able to run old topologies without a recompile (assuming no dependencies from storm were used, and I believe at least one config was removed, but you would have to check). When recompiling a new NotAuthorizedException can be thrown by some APIs that will need to be handled. From an end user perspective the big difference is in the dependencies, especially around shading of dependencies. This means that end users will want to recompile their topologies using at least 0.10.0, so that maven packages their dependencies in the uber jar correctly.
The REST API should still be backwards compatible.
- Bobby
On Monday, June 8, 2015 4:19 PM, P. Taylor Goetz <pt...@gmail.com> wrote:
I’m working on a blog post for the 0.10.0 release and plan on highlighting the main new features. There are a lot and I want to make sure I mention them. Below is what I have off the top of my head (in no particular order aside from the fist):
* Security (the big one)
* Support for future rolling upgrades
* Kafka integration improvements
* Hive integration
* MS Azure Eventhubs Integration
* Redis integration
* JDBC integration
* Partial key grouping
* Dependency conflict reduction
* Topology deployment and testing tool (flux)
* New logging framework
If you think I missed anything let me know. And if you think of any important points for any of the above that you’d like to see included, let me know.
I’d also like to include a list of changes that are not backward compatible. I’m not sure we need it for the announcement, but we should definitely add it to the documentation. We don’t currently track that information in JIRA (we should consider doing so), so if anyone can help with compiling a list, I’d appreciate it.
I will try to send out a draft for review tomorrow.
And most importantly, thanks and congratulations to everyone who contributed to this release, it’s a big one!
-Taylor