You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@maven.apache.org by Matthew Tordoff <ma...@hotmail.com> on 2008/04/11 17:52:49 UTC
Highly Available Archiva
Hi all,
Does anyone have experience in setting up Archiva in a highly available manner? What is the best considered approach for doing this? Is it possible to replicate deployments to a server across all other servers in a cluster? Are there any options for automatic failover?
Any advice or pointers on this would be greatly appreciated.
Kind Regards,
Matt T
_________________________________________________________________
The next generation of Windows Live is here
http://www.windowslive.co.uk/get-live
Re: Highly Available Archiva
Posted by Joshua ChaitinPollak <jp...@kivasystems.com>.
You could try LinuxHA, or a commercial solution like ServiceGuard or
SteelEye.
On Apr 11, 2008, at 11:52 AM, Matthew Tordoff wrote:
>
> Hi all,
>
> Does anyone have experience in setting up Archiva in a highly
> available manner? What is the best considered approach for doing
> this? Is it possible to replicate deployments to a server across all
> other servers in a cluster? Are there any options for automatic
> failover?
>
> Any advice or pointers on this would be greatly appreciated.
>
> Kind Regards,
>
> Matt T
>
> _________________________________________________________________
> The next generation of Windows Live is here
> http://www.windowslive.co.uk/get-live
--
Joshua ChaitinPollak | Software Engineer
Kiva Systems, Inc., 225 Wildwood Ave, Woburn, MA 01970
Re: Highly Available Archiva
Posted by Wendy Smoak <ws...@gmail.com>.
On Fri, Apr 11, 2008 at 9:39 PM, Nick Stolwijk <ni...@planet.nl> wrote:
> I know Archiva is run on a JCR (JackRabbit) repository. I've setup a
> Jackrabbit cluster on a clustered oracle database. (So twice clustering, one
> for the repo, one for the db). I think it shouldn't be too hard to implement
> this with Archiva also.
I believe Artifactory is the one that uses JCR behind the scenes.
For Archiva I use a setup similar to the one Jason described -- shared
filesystem with multiple servers behind a VIP. (You'll need to
configure separate index and database directories for each server as
those shouldn't be shared.)
--
Wendy
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
For additional commands, e-mail: users-help@maven.apache.org
Re: Highly Available Archiva
Posted by Jason van Zyl <ja...@maven.org>.
On 18-Apr-08, at 1:30 AM, Yoav Landman wrote:
>
> I guess the above approach is fine if you have basic requirements
> from a
> repository manager and are willing to go for one that:
>
> - Treats artifacts as just files, discarding any extra metadata that
> cannot
> be maintained using the file system, such as: who deployed an
> artifact, how
> many times it has been downloaded, inter-artifact dependencies etc.
We don't discard any artifact metadata. That doesn't need to be stored
in a database.
>
> - Does not need security and therefore does not have to deal with
> storage of
> ACLs.
Having security is not predicated on using a database, and we'll
demonstrate that.
>
> - Does not need atomic/transactional deployments and/or concurrency
> controls
> and willing to live with the chance of being left with half baked
> deployments.
Also not true. We'll also demonstrate this, but there's nothing you
can do right now because the problem is on the Maven side where there
is no atomic semantics so it's a two pronged approach for correct
operation. We're working on the Maven side and the repository manager
side to make sure it's simple, file-based, and safe.
>
>
> BTW, the latest Artifactory achieves HA by using *only* DB
> replication (not
> a combination of DB and FS). HA and ACID usually go together and
> Artifactory
> is doing it.
>
> Yoav Landman
> The Artifactory Project Lead
>
>
> Jason van Zyl-2 wrote:
>>
>> It's twice as hard then simply replicating a file system. Try in a
>> large organization getting DBAs and the sysadmins synced up. It's
>> just
>> easier replicating the filesystem. Less moving parts equals better.
>>
>> On 11-Apr-08, at 12:39 PM, Nick Stolwijk wrote:
>>> I know Archiva is run on a JCR (JackRabbit) repository. I've setup a
>>> Jackrabbit cluster on a clustered oracle database. (So twice
>>> clustering, one for the repo, one for the db). I think it shouldn't
>>> be too hard to implement this with Archiva also.
>>>
>>> Hth,
>>>
>>> Nick S.
>>>
>>> Jason van Zyl wrote:
>>>> Not Archiva but Nexus where the disk uses Raid 5 which is then a
>>>> network mount. The data and artifacts are shared between two
>>>> instances of Nexus and they sit behind a VIP. If the primary goes
>>>> down then the VIP flips over to the second instance that's running.
>>>> With Nexus it's simple disk replication so you could do lower tech
>>>> solutions using rsync, or not so low tech as there are tools that
>>>> do disk replication at near wire speeds. But the VIP approach has
>>>> worked for us.
>>>>
>>>> On 11-Apr-08, at 8:52 AM, Matthew Tordoff wrote:
>>>>>
>>>>> Hi all,
>>>>>
>>>>> Does anyone have experience in setting up Archiva in a highly
>>>>> available manner? What is the best considered approach for doing
>>>>> this? Is it possible to replicate deployments to a server across
>>>>> all other servers in a cluster? Are there any options for
>>>>> automatic failover?
>>>>>
>>>>> Any advice or pointers on this would be greatly appreciated.
>>>>>
>>>>> Kind Regards,
>>>>>
>>>>> Matt T
>>>>>
>>>>> _________________________________________________________________
>>>>> The next generation of Windows Live is here
>>>>> http://www.windowslive.co.uk/get-live
>>>>
>>>> Thanks,
>>>>
>>>> Jason
>>>>
>>>> ----------------------------------------------------------
>>>> Jason van Zyl
>>>> Founder, Apache Maven
>>>> jason at sonatype dot com
>>>> ----------------------------------------------------------
>>>>
>>>> believe nothing, no matter where you read it,
>>>> or who has said it,
>>>> not even if i have said it,
>>>> unless it agrees with your own reason
>>>> and your own common sense.
>>>>
>>>> -- Buddha
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>>>> For additional commands, e-mail: users-help@maven.apache.org
>>>>
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: users-help@maven.apache.org
>>>
>>
>> Thanks,
>>
>> Jason
>>
>> ----------------------------------------------------------
>> Jason van Zyl
>> Founder, Apache Maven
>> jason at sonatype dot com
>> ----------------------------------------------------------
>>
>> You are never dedicated to something you have complete confidence in.
>> No one is fanatically shouting that the sun is going to rise
>> tomorrow.
>> They know it is going to rise tomorrow. When people are fanatically
>> dedicated to political or religious faiths or any other kind of
>> dogmas or goals, it's always because these dogmas or
>> goals are in doubt.
>>
>> -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance
>>
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>> For additional commands, e-mail: users-help@maven.apache.org
>>
>>
>>
>
> --
> View this message in context: http://www.nabble.com/Highly-Available-Archiva-tp16635891s177p16762084.html
> Sent from the Maven - Users mailing list archive at Nabble.com.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
> For additional commands, e-mail: users-help@maven.apache.org
>
Thanks,
Jason
----------------------------------------------------------
Jason van Zyl
Founder, Apache Maven
jason at sonatype dot com
----------------------------------------------------------
In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
-- Jacques Ellul, The Technological Society
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
For additional commands, e-mail: users-help@maven.apache.org
Re: Highly Available Archiva
Posted by Yoav Landman <yl...@gmail.com>.
I guess the above approach is fine if you have basic requirements from a
repository manager and are willing to go for one that:
- Treats artifacts as just files, discarding any extra metadata that cannot
be maintained using the file system, such as: who deployed an artifact, how
many times it has been downloaded, inter-artifact dependencies etc.
- Does not need security and therefore does not have to deal with storage of
ACLs.
- Does not need atomic/transactional deployments and/or concurrency controls
and willing to live with the chance of being left with half baked
deployments.
BTW, the latest Artifactory achieves HA by using *only* DB replication (not
a combination of DB and FS). HA and ACID usually go together and Artifactory
is doing it.
Yoav Landman
The Artifactory Project Lead
Jason van Zyl-2 wrote:
>
> It's twice as hard then simply replicating a file system. Try in a
> large organization getting DBAs and the sysadmins synced up. It's just
> easier replicating the filesystem. Less moving parts equals better.
>
> On 11-Apr-08, at 12:39 PM, Nick Stolwijk wrote:
>> I know Archiva is run on a JCR (JackRabbit) repository. I've setup a
>> Jackrabbit cluster on a clustered oracle database. (So twice
>> clustering, one for the repo, one for the db). I think it shouldn't
>> be too hard to implement this with Archiva also.
>>
>> Hth,
>>
>> Nick S.
>>
>> Jason van Zyl wrote:
>>> Not Archiva but Nexus where the disk uses Raid 5 which is then a
>>> network mount. The data and artifacts are shared between two
>>> instances of Nexus and they sit behind a VIP. If the primary goes
>>> down then the VIP flips over to the second instance that's running.
>>> With Nexus it's simple disk replication so you could do lower tech
>>> solutions using rsync, or not so low tech as there are tools that
>>> do disk replication at near wire speeds. But the VIP approach has
>>> worked for us.
>>>
>>> On 11-Apr-08, at 8:52 AM, Matthew Tordoff wrote:
>>>>
>>>> Hi all,
>>>>
>>>> Does anyone have experience in setting up Archiva in a highly
>>>> available manner? What is the best considered approach for doing
>>>> this? Is it possible to replicate deployments to a server across
>>>> all other servers in a cluster? Are there any options for
>>>> automatic failover?
>>>>
>>>> Any advice or pointers on this would be greatly appreciated.
>>>>
>>>> Kind Regards,
>>>>
>>>> Matt T
>>>>
>>>> _________________________________________________________________
>>>> The next generation of Windows Live is here
>>>> http://www.windowslive.co.uk/get-live
>>>
>>> Thanks,
>>>
>>> Jason
>>>
>>> ----------------------------------------------------------
>>> Jason van Zyl
>>> Founder, Apache Maven
>>> jason at sonatype dot com
>>> ----------------------------------------------------------
>>>
>>> believe nothing, no matter where you read it,
>>> or who has said it,
>>> not even if i have said it,
>>> unless it agrees with your own reason
>>> and your own common sense.
>>>
>>> -- Buddha
>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>>> For additional commands, e-mail: users-help@maven.apache.org
>>>
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>> For additional commands, e-mail: users-help@maven.apache.org
>>
>
> Thanks,
>
> Jason
>
> ----------------------------------------------------------
> Jason van Zyl
> Founder, Apache Maven
> jason at sonatype dot com
> ----------------------------------------------------------
>
> You are never dedicated to something you have complete confidence in.
> No one is fanatically shouting that the sun is going to rise tomorrow.
> They know it is going to rise tomorrow. When people are fanatically
> dedicated to political or religious faiths or any other kind of
> dogmas or goals, it's always because these dogmas or
> goals are in doubt.
>
> -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
> For additional commands, e-mail: users-help@maven.apache.org
>
>
>
--
View this message in context: http://www.nabble.com/Highly-Available-Archiva-tp16635891s177p16762084.html
Sent from the Maven - Users mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
For additional commands, e-mail: users-help@maven.apache.org
Re: Highly Available Archiva
Posted by Jason van Zyl <ja...@maven.org>.
It's twice as hard then simply replicating a file system. Try in a
large organization getting DBAs and the sysadmins synced up. It's just
easier replicating the filesystem. Less moving parts equals better.
On 11-Apr-08, at 12:39 PM, Nick Stolwijk wrote:
> I know Archiva is run on a JCR (JackRabbit) repository. I've setup a
> Jackrabbit cluster on a clustered oracle database. (So twice
> clustering, one for the repo, one for the db). I think it shouldn't
> be too hard to implement this with Archiva also.
>
> Hth,
>
> Nick S.
>
> Jason van Zyl wrote:
>> Not Archiva but Nexus where the disk uses Raid 5 which is then a
>> network mount. The data and artifacts are shared between two
>> instances of Nexus and they sit behind a VIP. If the primary goes
>> down then the VIP flips over to the second instance that's running.
>> With Nexus it's simple disk replication so you could do lower tech
>> solutions using rsync, or not so low tech as there are tools that
>> do disk replication at near wire speeds. But the VIP approach has
>> worked for us.
>>
>> On 11-Apr-08, at 8:52 AM, Matthew Tordoff wrote:
>>>
>>> Hi all,
>>>
>>> Does anyone have experience in setting up Archiva in a highly
>>> available manner? What is the best considered approach for doing
>>> this? Is it possible to replicate deployments to a server across
>>> all other servers in a cluster? Are there any options for
>>> automatic failover?
>>>
>>> Any advice or pointers on this would be greatly appreciated.
>>>
>>> Kind Regards,
>>>
>>> Matt T
>>>
>>> _________________________________________________________________
>>> The next generation of Windows Live is here
>>> http://www.windowslive.co.uk/get-live
>>
>> Thanks,
>>
>> Jason
>>
>> ----------------------------------------------------------
>> Jason van Zyl
>> Founder, Apache Maven
>> jason at sonatype dot com
>> ----------------------------------------------------------
>>
>> believe nothing, no matter where you read it,
>> or who has said it,
>> not even if i have said it,
>> unless it agrees with your own reason
>> and your own common sense.
>>
>> -- Buddha
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
>> For additional commands, e-mail: users-help@maven.apache.org
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
> For additional commands, e-mail: users-help@maven.apache.org
>
Thanks,
Jason
----------------------------------------------------------
Jason van Zyl
Founder, Apache Maven
jason at sonatype dot com
----------------------------------------------------------
You are never dedicated to something you have complete confidence in.
No one is fanatically shouting that the sun is going to rise tomorrow.
They know it is going to rise tomorrow. When people are fanatically
dedicated to political or religious faiths or any other kind of
dogmas or goals, it's always because these dogmas or
goals are in doubt.
-- Robert Pirzig, Zen and the Art of Motorcycle Maintenance
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
For additional commands, e-mail: users-help@maven.apache.org
Re: Highly Available Archiva
Posted by Nick Stolwijk <ni...@planet.nl>.
I know Archiva is run on a JCR (JackRabbit) repository. I've setup a
Jackrabbit cluster on a clustered oracle database. (So twice clustering,
one for the repo, one for the db). I think it shouldn't be too hard to
implement this with Archiva also.
Hth,
Nick S.
Jason van Zyl wrote:
> Not Archiva but Nexus where the disk uses Raid 5 which is then a
> network mount. The data and artifacts are shared between two instances
> of Nexus and they sit behind a VIP. If the primary goes down then the
> VIP flips over to the second instance that's running. With Nexus it's
> simple disk replication so you could do lower tech solutions using
> rsync, or not so low tech as there are tools that do disk replication
> at near wire speeds. But the VIP approach has worked for us.
>
> On 11-Apr-08, at 8:52 AM, Matthew Tordoff wrote:
>>
>> Hi all,
>>
>> Does anyone have experience in setting up Archiva in a highly
>> available manner? What is the best considered approach for doing
>> this? Is it possible to replicate deployments to a server across all
>> other servers in a cluster? Are there any options for automatic
>> failover?
>>
>> Any advice or pointers on this would be greatly appreciated.
>>
>> Kind Regards,
>>
>> Matt T
>>
>> _________________________________________________________________
>> The next generation of Windows Live is here
>> http://www.windowslive.co.uk/get-live
>
> Thanks,
>
> Jason
>
> ----------------------------------------------------------
> Jason van Zyl
> Founder, Apache Maven
> jason at sonatype dot com
> ----------------------------------------------------------
>
> believe nothing, no matter where you read it,
> or who has said it,
> not even if i have said it,
> unless it agrees with your own reason
> and your own common sense.
>
> -- Buddha
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
> For additional commands, e-mail: users-help@maven.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
For additional commands, e-mail: users-help@maven.apache.org
Re: Highly Available Archiva
Posted by Jason van Zyl <ja...@maven.org>.
Not Archiva but Nexus where the disk uses Raid 5 which is then a
network mount. The data and artifacts are shared between two instances
of Nexus and they sit behind a VIP. If the primary goes down then the
VIP flips over to the second instance that's running. With Nexus it's
simple disk replication so you could do lower tech solutions using
rsync, or not so low tech as there are tools that do disk replication
at near wire speeds. But the VIP approach has worked for us.
On 11-Apr-08, at 8:52 AM, Matthew Tordoff wrote:
>
> Hi all,
>
> Does anyone have experience in setting up Archiva in a highly
> available manner? What is the best considered approach for doing
> this? Is it possible to replicate deployments to a server across all
> other servers in a cluster? Are there any options for automatic
> failover?
>
> Any advice or pointers on this would be greatly appreciated.
>
> Kind Regards,
>
> Matt T
>
> _________________________________________________________________
> The next generation of Windows Live is here
> http://www.windowslive.co.uk/get-live
Thanks,
Jason
----------------------------------------------------------
Jason van Zyl
Founder, Apache Maven
jason at sonatype dot com
----------------------------------------------------------
believe nothing, no matter where you read it,
or who has said it,
not even if i have said it,
unless it agrees with your own reason
and your own common sense.
-- Buddha
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@maven.apache.org
For additional commands, e-mail: users-help@maven.apache.org