Hi Yeah there could be a default implementation using JPA such as camel-bam does. And having "hooks" or whatever for others to extend/integrate a different data store.
But I guess it would be nice for Camel to know that the message is "claimed checked" so its body is *null* or whatever. And that if needed it needs to get the body by get it from X. We should be aware that Camel doesn't do this automatically so if you use a logger/tracer etc. and it outputs: getBody() it will not claim the body all the time. Med venlig hilsen Claus Ibsen ...................................... Silverbullet Skovsgårdsvænget 21 8362 Hørning Tlf. +45 2962 7576 Web: www.silverbullet.dk -----Original Message----- From: Martin Gilday [mailto:[EMAIL PROTECTED] Sent: 23. september 2008 11:54 To: camel-user@activemq.apache.org Subject: RE: Sending an ActiveMQ BlobMessage via camel I think in some ways the claim check pattern varies a lot for each use case. The developer generally needs to be in control of what is persisted and how. It is failry straightforward to just persist to a datasource and then start passing around an ID instead. Maybe there is some scope for having a hook for the user to provide what should be persisted and have the result of that callback be passed as the body to the next part of the route. Maybe there is the possibility of reusing some of the JPA component to do this. Of course you can use other data sources other than a database with the claim check idea, so that would need to be considered. ----- Original message ----- From: "Claus Ibsen" <[EMAIL PROTECTED]> To: camel-user@activemq.apache.org Date: Tue, 23 Sep 2008 11:39:19 +0200 Subject: RE: Sending an ActiveMQ BlobMessage via camel Hi I think for startes we should have a ticket in JIRA so we have the request that end-users need this pattern. Also it will not be lost as old messages in this forum tend to be. Then other end-users and the other camel riders can write feedback on this issue as well. Then I think you should build whatever solution that works for you in your situation and then we can use that as a starting point to integrate into Camel and make the core more "Camelish" and more generic if needed/possible. As Camel is shipped with ActiveMQ and they are much integrated I can envision that it would be a bit easier to integrated claim check with BlobMessages into ActiveMQ and Camel. Med venlig hilsen Claus Ibsen ...................................... Silverbullet Skovsgårdsvænget 21 8362 Hørning Tlf. +45 2962 7576 Web: www.silverbullet.dk -----Original Message----- From: jayson.minard [mailto:[EMAIL PROTECTED] Sent: 23. september 2008 11:21 To: camel-user@activemq.apache.org Subject: RE: Sending an ActiveMQ BlobMessage via camel Claus, Sure. What is the philosophy of the project for handling something such as this which is built into one underlying system, but probably not many of similar underlying systems? Implement it with a generic approach + a specific ActiveMQ approach, or a generic approach only? --j Claus Ibsen wrote: > > Hi > > Ah thanks. > > The claim-check EIP pattern is not implemented in Camel. > > We love contributions, so if you build this we can integrate it into > Camel: > http://activemq.apache.org/camel/contributing.html > > For starters you can create a ticket for the claim check EIP pattern, so > its not forgotten. > > > Med venlig hilsen > > Claus Ibsen > ...................................... > Silverbullet > Skovsgårdsvænget 21 > 8362 Hørning > Tlf. +45 2962 7576 > Web: www.silverbullet.dk > > -----Original Message----- > From: jayson.minard [mailto:[EMAIL PROTECTED] > Sent: 23. september 2008 10:56 > To: camel-user@activemq.apache.org > Subject: RE: Sending an ActiveMQ BlobMessage via camel > > > A BlobMessage in ActiveMQ allows the body to be stored out-of-band. > Basically it is a claim-check pattern with the body being stored on FTP, > file system, HTTP or some other remote store while the message flies > through > the queues with a ticket to claim the body if ever requested. In this > model > you rarely request the body but want access to it when you need it. So it > is not a BytesMessage, but an ActiveMQ-ism. > > The http://activemq.apache.org/blob-messages.html ActiveMQ site has an > explanation . > > --j > > > > Claus Ibsen wrote: >> >> is BlobMessage an ActiveMQ type? >> >> What javax JMS Message type does it implement? javax.jmx.BytesMessage. >> >> I guess the fault could be that Camel doesn't regard it as a Bytes >> Message >> and want to convert it. >> >> I guess James would be the ideal candidate to answer this one. >> >> >> >> Med venlig hilsen >> >> Claus Ibsen >> ...................................... >> Silverbullet >> Skovsgårdsvænget 21 >> 8362 Hørning >> Tlf. +45 2962 7576 >> Web: www.silverbullet.dk >> >> -----Original Message----- >> From: jayson.minard [mailto:[EMAIL PROTECTED] >> Sent: 23. september 2008 10:34 >> To: camel-user@activemq.apache.org >> Subject: Sending an ActiveMQ BlobMessage via camel >> >> >> I haven't been able to find a full working sample of sending an ActiveMQ >> (5.1.0) BlobMessage via Camel (1.4.0). Has anyone accomplished this, and >> have a pointer or two? >> >> We are sending very large files around and our performance is hurting >> when >> the body of the message is passed between many route processors that >> never >> want the body to begin with. If we cannot use BlobMessage we'll need to >> write our own claim ticket approach... >> >> Thanks in advance for any comments! >> Cheers, >> -- Jayson >> -- >> View this message in context: >> http://www.nabble.com/Sending-an-ActiveMQ-BlobMessage-via-camel-tp19623413s22882p19623413.html >> Sent from the Camel - Users mailing list archive at Nabble.com. >> >> >> > > -- > View this message in context: > http://www.nabble.com/Sending-an-ActiveMQ-BlobMessage-via-camel-tp19623413s22882p19623771.html > Sent from the Camel - Users mailing list archive at Nabble.com. > > > -- View this message in context: http://www.nabble.com/Sending-an-ActiveMQ-BlobMessage-via-camel-tp19623413s22882p19624099.html Sent from the Camel - Users mailing list archive at Nabble.com.
