6:02 p.m.
I just submitted a new PR for hawkular services:
https://github.com/hawkular/hawkular-services/pull/22
This re-introduces the ability to build Hawkular Services with embedded Cassandra if you
want.
This will ONLY embed Cassandra IF you build with the special Maven profile -Pembeddedc.
If you never build with that profile, Hawkular Services will build as it does now (that
is, requires you to install and start your own Cassandra separately).
So this will allow developers to continue doing what they've been doing (that is, no
need to run a standalone C*) if they want.
Again, the default build is NOT to embed Cassandra. You must explicitly tell Maven you
want to embed C* via -Pembeddedc to get this feature.
6:08 p.m.
On 6/20/2016 12:02 PM, John Mazzitelli wrote:
I just submitted a new PR for hawkular services:
https://github.com/hawkular/hawkular-services/pull/22
This re-introduces the ability to build Hawkular Services with embedded Cassandra if you
want.
This will ONLY embed Cassandra IF you build with the special Maven profile -Pembeddedc.
I'm +1 on this option and external c* being the default. But -1 on that
profile name, can you make it something obvious like -Pembedded-cassandra
If you never build with that profile, Hawkular Services will build as it does now (that
is, requires you to install and start your own Cassandra separately).
So this will allow developers to continue doing what they've been doing (that is, no
need to run a standalone C*) if they want.
Again, the default build is NOT to embed Cassandra. You must explicitly tell Maven you
want to embed C* via -Pembeddedc to get this feature.
_______________________________________________
hawkular-dev mailing list
hawkular-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev
6:17 p.m.
I'm +1 on this option and external c* being the default. But -1
on that
profile name, can you make it something obvious like -Pembedded-cassandra
I tried limiting the number of characters I'm gonna have to type every time I build
this now :)
How about something smaller than "embedded-cassandra"?? :)
7:04 p.m.
-Pmazz
-Pembedded
-Pdev-cassandra
I dunno, leave it as -Pembeddedc if you like.
On 6/20/2016 12:17 PM, John Mazzitelli wrote:
> I'm +1 on this option and external c* being the default. But
-1 on that
> profile name, can you make it something obvious like -Pembedded-cassandra
I tried limiting the number of characters I'm gonna have to type every time I build
this now :)
How about something smaller than "embedded-cassandra"?? :)
_______________________________________________
hawkular-dev mailing list
hawkular-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev
9:57 p.m.
Mazz,
I have nothing against your PR in particular, but I'm against this
feature in principle: it's more code that we have to take care of for
"development only". And as it's more code, it also means more bugs.
Perhaps not now, but in a week or two (or once we have a change in the
XML templates).
To me, requiring the developer to install Cassandra is similar to
requiring a developer to install MariaDB or PostgreSQL and that's
acceptable, specially if we don't have to maintain code that is to be
used only for development purposes.
On a Fedora machine, Cassandra can be installed and started by:
# dnf config-manager --add-repo=http://rpm.datastax.com/community
# dnf install dsc30
# systemctl start cassandra
( plus, there might some required changes related to networking in
/etc/cassandra/conf/cassandra.yaml )
That said, I suggest -Paio (all-in-one) for the profile name :)
- Juca.
On 20.06.2016 18:02, John Mazzitelli wrote:
I just submitted a new PR for hawkular services:
https://github.com/hawkular/hawkular-services/pull/22
This re-introduces the ability to build Hawkular Services with embedded Cassandra if you
want.
This will ONLY embed Cassandra IF you build with the special Maven profile -Pembeddedc.
If you never build with that profile, Hawkular Services will build as it does now (that
is, requires you to install and start your own Cassandra separately).
So this will allow developers to continue doing what they've been doing (that is, no
need to run a standalone C*) if they want.
Again, the default build is NOT to embed Cassandra. You must explicitly tell Maven you
want to embed C* via -Pembeddedc to get this feature.
_______________________________________________
hawkular-dev mailing list
hawkular-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev
10:04 p.m.
I know that there has been reluctance to using ccm, but that really should be part of the
development environment. It can run on any platform that has python. You can easily manage
single- or multi-node clusters that run locally. I walked Jay through it in less than 5
minutes and had him up and running with these simple steps:
1. pip install ccm
2. ccm create test -v 3.7 -n 1 —nodes
3. Edit ~/.ccm/test/node1/conf/cassandra.yaml and set the start_rpc property to true
(needed for inventory)
4. ccm stop
5. ccm start
On Jun 20, 2016, at 3:57 PM, Juraci Paixão Kröhling
<jpkroehling(a)redhat.com> wrote:
Mazz,
I have nothing against your PR in particular, but I'm against this
feature in principle: it's more code that we have to take care of for
"development only". And as it's more code, it also means more bugs.
Perhaps not now, but in a week or two (or once we have a change in the
XML templates).
To me, requiring the developer to install Cassandra is similar to
requiring a developer to install MariaDB or PostgreSQL and that's
acceptable, specially if we don't have to maintain code that is to be
used only for development purposes.
On a Fedora machine, Cassandra can be installed and started by:
# dnf config-manager --add-repo=http://rpm.datastax.com/community
# dnf install dsc30
# systemctl start cassandra
( plus, there might some required changes related to networking in
/etc/cassandra/conf/cassandra.yaml )
That said, I suggest -Paio (all-in-one) for the profile name :)
- Juca.
On 20.06.2016 18:02, John Mazzitelli wrote:
> I just submitted a new PR for hawkular services:
>
> https://github.com/hawkular/hawkular-services/pull/22
>
> This re-introduces the ability to build Hawkular Services with embedded Cassandra if
you want.
>
> This will ONLY embed Cassandra IF you build with the special Maven profile
-Pembeddedc.
>
> If you never build with that profile, Hawkular Services will build as it does now
(that is, requires you to install and start your own Cassandra separately).
>
> So this will allow developers to continue doing what they've been doing (that is,
no need to run a standalone C*) if they want.
>
> Again, the default build is NOT to embed Cassandra. You must explicitly tell Maven
you want to embed C* via -Pembeddedc to get this feature.
> _______________________________________________
> hawkular-dev mailing list
> hawkular-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/hawkular-dev
>
_______________________________________________
hawkular-dev mailing list
hawkular-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev
10:43 p.m.
New subject: hawkular services with embedded c profile - need PR reviewed/merged
Honestly, I don't get it. We have a nice all-in-one way to run hawkular server. I
don't want to have to spend even 5 minutes learning how great ccm is, or how to
download/install/start C* on its own, when the feature exists that I can just build and
run.
We have a way to embed our datasource! That's awesome! Why are we afraid to use it or
support it - even from just a dev or demo standpoint?
BTW: We are still using embedded C* in our itests - so we are still in need to support
it.
Anyway, as I say, this is a completely separate, standalone maven profile. You never have
to use it - you don't even need to know it exists. Just ignore it if you want. But for
people who want it, its there. Win-win.
As for the fact it is more code, I will see if I can get the pom to generate its own
feature-branch xml file so we only have one. When that happens, there will be no real
extra config to maintain. The only thing extra will be the use of this hawkular.embeddedc
property - which really just mimics the hawkular.rest.url/password stuff.
11:39 p.m.
New subject: hawkular services with embedded c profile - need PR reviewed/merged
As for the fact it is more code, I will see if I can get the pom to
generate
its own feature-branch xml file so we only have one. When that happens,
there will be no real extra config to maintain. The only thing extra will be
the use of this hawkular.embeddedc property - which really just mimics the
hawkular.rest.url/password stuff.
This is done. No more duplicated feature pack build xml files. Back to just having the
one.
8:33 a.m.
On 20 Jun 2016, at 21:57, Juraci Paixão Kröhling wrote:
That said, I suggest -Paio (all-in-one) for the profile name :)
Which reminds me that we (Hi Stefan :-) are working on an AIO
download which then indeed includes embedded-Cassandra,
hawkular-services and explorer UI.
Perhaps that build could also satisfy that "C* for developers" need(?)
--
Reg. Adresse: Red Hat GmbH, Technopark II, Haus C,
Werner-von-Siemens-Ring 14, D-85630 Grasbrunn
Handelsregister: Amtsgericht München HRB 153243
Geschäftsführer: Charles Cachera, Michael Cunningham, Michael O'Neill,
Eric Shander
2:14 p.m.
----- Original Message -----
On 20 Jun 2016, at 21:57, Juraci Paixão Kröhling wrote:
> That said, I suggest -Paio (all-in-one) for the profile name :)
Which reminds me that we (Hi Stefan :-) are working on an AIO
download which then indeed includes embedded-Cassandra,
hawkular-services and explorer UI.
Perhaps that build could also satisfy that "C* for developers" need(?)
Which tells me we probably shouldn't call this maven profile "aio" since it
may be confused with that hawkular build (i.e. it really isn't all in one because it
won't have that explorer UI for one thing). This maven profile only includes embedded
C* and is part of the hawkular-services repo, not the hawkular repo.
And as we just saw with Austin's timely post, it is clear to me that we need to
provide this - let's make it as easy as possible for people to start doing development
on Hawkular. Why put more roadblocks in their way when we don't need to? It also makes
it easier on people like me who don't want to deal with it when I don't have to.
:)
2:45 p.m.
On 21 Jun 2016, at 14:14, John Mazzitelli wrote:
need to provide this - let's make it as easy as possible for
people to
start doing development on Hawkular. Why put more roadblocks in their
way when we don't need to? It also makes it easier on people like me
who don't want to deal with it when I don't have to. :)
For those just "consuming" it,
$ docker-compose docker-compose.yml
from attached file could be even easier :)
--
Reg. Adresse: Red Hat GmbH, Technopark II, Haus C,
Werner-von-Siemens-Ring 14, D-85630 Grasbrunn
Handelsregister: Amtsgericht München HRB 153243
Geschäftsführer: Charles Cachera, Michael Cunningham, Michael O'Neill,
Eric Shander
3:20 p.m.
On 21 Jun 2016, at 14:14, John Mazzitelli wrote:
> need to provide this - let's make it as easy as possible for people to
> start doing development on Hawkular. Why put more roadblocks in their
> way when we don't need to? It also makes it easier on people like me
> who don't want to deal with it when I don't have to. :)
For those just "consuming" it,
$ docker-compose docker-compose.yml
from attached file could be even easier :)
Is the file system mounted somewhere locally so its easy to read or tail the logs?
If hawkular-services can be built like this and give me everything:
1. mvn -Pdocker clean install
2. docker run ./target/hawkular-services-docker (or whatever the command is)
3. ...See the logs on the console and be able to read the log file as if its on a local
mount point...
Then that's good enough for me. Of course, that would probably be more code to have to
maintain, so that's bad :)
3:23 p.m.
I guess I should also ask - how long does it take to build that docker container? A
minute? Or a half-hour?
----- Original Message -----
> On 21 Jun 2016, at 14:14, John Mazzitelli wrote:
> > need to provide this - let's make it as easy as possible for people to
> > start doing development on Hawkular. Why put more roadblocks in their
> > way when we don't need to? It also makes it easier on people like me
> > who don't want to deal with it when I don't have to. :)
>
> For those just "consuming" it,
>
> $ docker-compose docker-compose.yml
> from attached file could be even easier :)
Is the file system mounted somewhere locally so its easy to read or tail the
logs?
If hawkular-services can be built like this and give me everything:
1. mvn -Pdocker clean install
2. docker run ./target/hawkular-services-docker (or whatever the command is)
3. ...See the logs on the console and be able to read the log file as if its
on a local mount point...
Then that's good enough for me. Of course, that would probably be more code
to have to maintain, so that's bad :)
_______________________________________________
hawkular-dev mailing list
hawkular-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev
3:30 p.m.
On 21 Jun 2016, at 15:23, John Mazzitelli wrote:
I guess I should also ask - how long does it take to build that
docker
container? A minute? Or a half-hour?
It will take the initial time to download the image(s) and then the
startup is as quick as directly calling bin/standalone.sh
--
Reg. Adresse: Red Hat GmbH, Technopark II, Haus C,
Werner-von-Siemens-Ring 14, D-85630 Grasbrunn
Handelsregister: Amtsgericht München HRB 153243
Geschäftsführer: Charles Cachera, Michael Cunningham, Michael O'Neill,
Eric Shander
3:39 p.m.
But what about "code-build-run" cycles as you are developing something? I am
assuming the docker container can be built right into the target/ directory of the
hawkular-services/dist directory (or I guess a new hawkular-services/docker location) and
I just run the docker container from there.
Why would you have to download the image after you just built it locally?
----- Original Message -----
On 21 Jun 2016, at 15:23, John Mazzitelli wrote:
> I guess I should also ask - how long does it take to build that docker
> container? A minute? Or a half-hour?
It will take the initial time to download the image(s) and then the
startup is as quick as directly calling bin/standalone.sh
--
Reg. Adresse: Red Hat GmbH, Technopark II, Haus C,
Werner-von-Siemens-Ring 14, D-85630 Grasbrunn
Handelsregister: Amtsgericht München HRB 153243
Geschäftsführer: Charles Cachera, Michael Cunningham, Michael O'Neill,
Eric Shander
_______________________________________________
hawkular-dev mailing list
hawkular-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/hawkular-dev
9:36 p.m.
On 21 Jun 2016, at 15:39, John Mazzitelli wrote:
But what about "code-build-run" cycles as you are
developing
something? I am assuming the docker container can be built right into
the target/ directory of the hawkular-services/dist directory (or I
guess a new hawkular-services/docker location) and I just run the
docker container from there.
Yes.
Why would you have to download the image after you just built it
locally?
This was more meant if you only want to consume the hawkular-services
server like in the scenario of working on the MiQ provider or writing
a new external agent.
For local server development it is probably best to just run the
server "bare metal". C* could still live in Docker though. But here
the advantage is certainly less.
--
Reg. Adresse: Red Hat GmbH, Technopark II, Haus C,
Werner-von-Siemens-Ring 14, D-85630 Grasbrunn
Handelsregister: Amtsgericht München HRB 153243
Geschäftsführer: Charles Cachera, Michael Cunningham, Michael O'Neill,
Eric Shander
10:29 p.m.
> Why would you have to download the image after you just built it
locally?
This was more meant if you only want to consume the hawkular-services
server like in the scenario of working on the MiQ provider or writing
a new external agent.
For local server development it is probably best to just run the
server "bare metal". C* could still live in Docker though. But here
the advantage is certainly less.
Hence, why I want -Pembeddedc in hawkular services build - no need for separate C* docker
container to run! :)
Am I going to be able to get this PR merged? Again, it is not the default, and no one even
has to know it exists if you don't want to use it - you never have to use it if you
love running C* externally. And the change isn't even large. There isn't a lot
going on under the covers here to support this.
https://github.com/hawkular/hawkular-services/pull/22
3:58 p.m.
On 21 Jun 2016, at 15:20, John Mazzitelli wrote:
Is the file system mounted somewhere locally so its easy to read or
tail the logs?
When you start docker with -it argument it will just forward the console
log of the server to stdout. This is of course not enough for
development
inside the server.
Right now, only the data directory is exposed, but that can be changed.
If hawkular-services can be built like this and give me everything:
1. mvn -Pdocker clean install
2. docker run ./target/hawkular-services-docker (or whatever the
command is)
3. ...See the logs on the console and be able to read the log file as
if its on a local mount point...
This mostly works like this in the docker-dist3 directory of my branch.
You do
$ mvn -Pdev install
$ cd docker-dist3
$ mvn docker:build
$ mvn docker:run
I am not saying this is a panacea, but certainly helpful to some
Then that's good enough for me. Of course, that would probably be
more
code to have to maintain, so that's bad :)
Well. That Docker route is getting some traction at the moment, so
it is certainly something we may want to keep going forward.
--
Reg. Adresse: Red Hat GmbH, Technopark II, Haus C,
Werner-von-Siemens-Ring 14, D-85630 Grasbrunn
Handelsregister: Amtsgericht München HRB 153243
Geschäftsführer: Charles Cachera, Michael Cunningham, Michael O'Neill,
Eric Shander
2872
days inactive
2873
days old
17 comments
5 participants
participants (5)
-
Heiko W.Rupp -
Jay Shaughnessy -
John Mazzitelli -
John Sanda -
Juraci Paixão Kröhling