Re: 4.3.0??
I had missed your message about RC1 as well, that's very exciting. I keep checking the Karaf page looking for 4.3.0 more and more frequently since I've been playing with R7 more and more. Can't wait to try it with Karaf. Thanks for all your hard work JB Ryan -- Sent from: http://karaf.922171.n3.nabble.com/Karaf-User-f930749.html
Re: Re: Karaf pax-logging-log4j2 bundle gets restarted - how to avoid?
I see this problem a lot when deploying my project in Karaf, for example, I was trying to allow installing different features independently depending on which capability someone might need, a feature that uses the Jackson libraries for example (either directly or via camel-jackson or one of the other myriads of projects that use jackson). If any features are already installed, this causes a cascading reload of just about everything. I guess it feels like the main karaf dependencies shouldn't be refreshed by default since it can cause strange stuff to happen. I'm sure it's not as simple as that, but it definitely does bug me as well that pax-logging-log4j2 would have a dependency on Jackson, Kafka, ZeroMQ, etc... Jackson is at least somewhat stomach-able as a dependency, but not ZeroMQ or Kafka. I'm not sure how pax-logging built, so I'm not sure if it's possible to control any of this, or if this is because of something like log4j being coupled to every library it possible can be coupled to (not the first library to do that sort of thing and unfortunately won't be the last) ;) For now I'm just trying to identify dependencies that feel like I can justify pre-installing in a custom distribution to try to get around this. It doesn't solve the whole problem but at least some of it. Ryan -- Sent from: http://karaf.922171.n3.nabble.com/Karaf-User-f930749.html
Building Karaf
I just wanted to see if anyone was having issues with the latest Karaf trunk. When I try to build it, it fails on the tests in the main subproject (the output from the tests is posted below), and when I use the latest snapshot in my other integration tests (which used to run), the RemoteBundleContextClient cannot be retrieved (it says it cannot connect to the 21414 port). I haven't modified any of the source I'm building from, and it used to always build fine for me. Does this sound like anything anyone has ran into before? Thanks for any help! --- T E S T S --- Running org.apache.karaf.main.lock.DefaultJDBCLockIntegrationTest Tests run: 1, Failures: 0, Errors: 0, Skipped: 1, Time elapsed: 0.035 sec Running org.apache.karaf.main.lock.DefaultJDBCLockTest Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock schemaExists SEVERE: Error testing for db table: java.lang.NullPointerException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock createSchema SEVERE: Could not create schema: java.lang.IllegalStateException: missing behavior definition for the preceding method call: Connection.getMetaData() Usage is: expect(a.foo()).andXXX() Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock isAlive SEVERE: Lost lock! Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock isAlive SEVERE: Lost lock! Tests run: 14, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.313 sec Running org.apache.karaf.main.lock.DerbyJDBCLockIntegrationTest Tests run: 1, Failures: 0, Errors: 0, Skipped: 1, Time elapsed: 0.004 sec Running org.apache.karaf.main.lock.DerbyJDBCLockTest Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock schemaExists SEVERE: Error testing for db table: java.lang.NullPointerException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock createSchema SEVERE: Could not create schema: java.lang.IllegalStateException: missing behavior definition for the preceding method call: Connection.getMetaData() Usage is: expect(a.foo()).andXXX() Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock isAlive SEVERE: Lost lock! Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock isAlive SEVERE: Lost lock! Tests run: 14, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.08 sec Running org.apache.karaf.main.lock.MySQLJDBCLockIntegrationTest Tests run: 1, Failures: 0, Errors: 0, Skipped: 1, Time elapsed: 0.003 sec Running org.apache.karaf.main.lock.MySQLJDBCLockTest Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock schemaExists SEVERE: Error testing for db table: java.lang.NullPointerException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock createSchema SEVERE: Could not create schema: java.lang.IllegalStateException: missing behavior definition for the preceding method call: Connection.getMetaData() Usage is: expect(a.foo()).andXXX() Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock updateLock WARNING: Failed to update database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock isAlive SEVERE: Lost lock! Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock isAlive SEVERE: Lost lock! Tests run: 14, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.048 sec Running org.apache.karaf.main.lock.OracleJDBCLockIntegrationTest Tests run: 1, Failures: 0, Errors: 0, Skipped: 1, Time elapsed: 0.007 sec Running org.apache.karaf.main.lock.OracleJDBCLockTest Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock aquireLock WARNING: Failed to acquire database lock: java.sql.SQLException Dec 01, 2013 12:15:26 PM org.apache.karaf.main.lock.DefaultJDBCLock aquireLock WARNING:
Re: Cellar 3.0.0-SNAPSHOT
Awesome, thanks!! :) -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-3-0-0-SNAPSHOT-tp4028863p4029075.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Cellar 3.0.0-SNAPSHOT
Good call! I just assumed it would use whatever followed as the command and just didn't care about quotes. So I'm guessing by the s at the end of features? I think I almost finished up the work I wanted to do on that fork. One quick fix to make and check in which should fix a random error I see with Hazelcast 3.0. I think think the blueprint files in 3.0 might have had some cyclical type dependencies and a lot of unnecessary service references. It feels much faster to me. I slimmed down a lot of stuff and have been diffing against the latest code, luckily for me it hasn't changed much, to make sure I'm not losing anything that is needed by accident. -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-3-0-0-SNAPSHOT-tp4028863p4029076.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Cellar 3.0.0-SNAPSHOT
Will -p be added to 3.0.0 as well? Otherwise unit tests using 3.0.0 that create and connect to a child instance won't be able to in an integration test. Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-3-0-0-SNAPSHOT-tp4028863p4029056.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Cellar 3.0.0-SNAPSHOT
Ok, cool, I figured but just wanted to make sure. Is there an rough estimate on on that feature? Just trying to plan ahead :) Thanks! -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-3-0-0-SNAPSHOT-tp4028863p4029058.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Features XSD
Awesome thanks! -- View this message in context: http://karaf.922171.n3.nabble.com/Features-XSD-tp4029006p4029046.html Sent from the Karaf - User mailing list archive at Nabble.com.
RE: How to access data folder?
It's funny because a lot of example services seem to use the data directory
to store things. I'm not one to say whether you should or shouldn't (though
in my opinion, I'm not sure where else you would store the database that you
know will exist), here is how you do it, you need to add the blueprint
extension namespace, a snippet of the blueprint tag is below:
blueprint xmlns=http://www.osgi.org/xmlns/blueprint/v1.0.0;
xmlns:cm=http://aries.apache.org/blueprint/xmlns/blueprint-cm/v1.1.0;
xmlns:ext=http://aries.apache.org/blueprint/xmlns/blueprint-ext/v1.0.0;
Next, add the following element:
ext:property-placeholder /
This will allow you to use to use karaf properties, such as the
${karaf.data} in a property tag like this:
property name=dbDirectory value=${karaf.data}/db/
--
View this message in context:
http://karaf.922171.n3.nabble.com/How-to-access-data-folder-tp4028895p4029044.html
Sent from the Karaf - User mailing list archive at Nabble.com.
Re: HazelcastFactory in cellar blueprint xml
Gotcha the docs didn't make it super clear either. I saw the thing about partitioning though ... makes me wonder how much overhead there is to having more than one hazelcast instance... seems like it might be a bit. -- View this message in context: http://karaf.922171.n3.nabble.com/HazelcastFactory-in-cellar-blueprint-xml-tp4028850p4028884.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Cellar 3.0.0-SNAPSHOT
I tried -p and it complains it's not a valid parameter, is there anyway to turn off security temporarily so I can see if I was able to fix 3.0.0? I'm still crossing my fingers, but I'm thinking I might have. -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-3-0-0-SNAPSHOT-tp4028863p4028885.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Cellar 2.3.0/3.0.x versions
Yep, I'm waiting to have something complete to sent but jb usually beats me on finishing since there is some complexity there I'm not familiar with. My goal is to get a big bug to jb before he can get to it so I can give him a hand. I love Karaf and what it's trying to accomplish, especially cell so I'm dying to help however I can. I think I found some invaluable for JB that I'm going to communicate to him about cellar. Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-2-3-0-3-0-x-versions-tp4028838p4028859.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Cellar 2.3.0/3.0.x versions
Yes, desperately. Karaf has worked so perfectly that I thought I'd just have to pull in cellar and I'd be good, the last few months I just haven't had any luck with Cellar but architecturally it's exactly what I want. When are you typically available? I'm pushing this hard and I just need clustering to work so I can finish the last bit of code. I gave up on cellar 2.3.0 due to an issue with felix having resource locking contention (it might be in my 2.3.0 cellar fork for rmoquin, I reordered your features.xml because I think your start orded is really out of order due to a single class. I'll file some bug reports. I decided to work on 3.0.0 because I seem to be having more luck. I've almost cracked to hazelcast classloading issues that causes all the TCCL blocks you use all over (which I'm convinced might be stomping on each other). I removed all of them in my 3.00 fork and I wrap the bundle object in a classloader subclass and call setClassloader on the hazelcast config with it. The only issue, is that the bundle disappears from the custom classloader probably when the instance is used in another bundle... so if i just reset it to the current bundle... then classloading works i have one or two other potential options as well.. take a look at my commits in cellar-hazelcast under rmoquin in github and I'll be that will help you a lot. If you can help me with advice on how I can finish fixing 3.0.0 I'll do the work and submit it back. I started programming in C when I was 11, so I'm more than capable with a little help. Thanks for all your hardwork. Rya Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-2-3-0-3-0-x-versions-tp4028838p4028860.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: ConfigAdmin services
Thanks JB, it didn't actually look like there was a link and so it made me wonder. Thanks for clearing up :) -- View this message in context: http://karaf.922171.n3.nabble.com/ConfigAdmin-services-tp4028830p4028834.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: startup vs. boot features
Ok, thanks for clarifying... I started to realize after a while, that nothing ever worked when I put it in as a startup features and I started getting a sneaking suspicion it's not what I thought it was. -- View this message in context: http://karaf.922171.n3.nabble.com/startup-vs-boot-features-tp4028826p4028836.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Clojure and OSGi/Karaf
You'll probably want to take a look at my fork of that project, https://github.com/rmoquin/clojure.osgi. It took me forever to figure out how to initially use it, and I made a bunch of enhancements to it. Basically there is an OSGi service you can use to execute something with a Clojure classloading context ... it works pretty well, though I'm sure there are some other additions it could use. Clojure stacktraces are immensly ugly and I found that in 1.4 that when an exception occurs, sometimes they swallow it (I think there was a comment about some exception hack they were using) and only print out a generic message that gives no indication of the error. I'm not sure if the osgi version of the clojure 1.4 lib has a modification or two to deal with it, or if I did something else. If you look at some of my other code, you can see how to use that clojure.osgi lib. The single classloader issue in Clojure really burned a lot of my time trying to deal with running a large amount of clojure code in Java. Oh well, hopefully it helps you out. Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Clojure-and-OSGi-Karaf-tp4028687p4028837.html Sent from the Karaf - User mailing list archive at Nabble.com.
Cellar 2.3.0/3.0.x versions
Hey JB, I didn't know if it was appropriate to ask here, but I was curious if you had fixes in the pipeline for cellar 3.0.x in particular. I'm asking because I checked out the SVN code (I know you guys don't consider github official yet) and was able to resolve a number of issues in head which actually make most of the integration tests pass (they actually fly pretty fast). I am trying to prepare a demo using Cellar and have been having issues with both versions (obviously) and was hoping there was a chance that some of the stuff I did might be useful to you. I know you where a lot of hats and your time is limited. I'm fascinated by cellar, so it's been interesting troubleshooting it. Let me know, I can submit a patch in a jira for you, I just don't want to overload you with a bunch of stuff that you've already squashed. :) Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Cellar-2-3-0-3-0-x-versions-tp4028838.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: kar files and external dependencies
I got so spoiled with karaf 3.0.0 and it's assembly project type that moving back to creating a distro with 2.3.x has been painful. I am missing customizing something correctly since everything seems to be in the build (though it ends up almost 15 megs bigger than when using the karaf-assembly project type, not sure if it really should be or not or if I have another goof). Unfortunately though, I can't use 3.0.0 of I want to use cellar because at runtime it goes crazy saying the groupmanager is missing, hazelcast shuts down and the container shortly then runs put of memory. My couple small int tests also have cellar load though I'm not taking advantage of it in the tests, but doesn't seem to go crazy on me maybe not fully initialized yet.) It looked like JB had fixed a lot of the issues that had been there so I'm surprised by how it's behaving. I have everything set correctly though I'm wondering if there is a startup order issue. Hopefully JB can confirm (sorry I keep asking about it's status but I really like Karaf and Cellar). I think some of my problem with 2.3.x and a custom distro is that I'm trying to create both a zip and gz and it seems like the maven life cycle might be executing twice. I'm not sure if the example could be updated to show creating both correctly or it really is easy and I'm just overlooking something. Thanks for all your hard work! Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/kar-files-and-external-dependencies-tp4028764p4028811.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Debugging my weird console issue
Ok, so I found out one interesting thing. I decided to just see what happened if I used putty to SSH into the console. When I did and hit tab, then it asks me if I want to see the list of commands. So it appears this problem is localized to when I run the karaf command and use the local console which is strange.. still not sure what could cause a difference between RC1 and the current snapshots that would cause that behavior only on that local console instance. -- View this message in context: http://karaf.922171.n3.nabble.com/Debugging-my-weird-console-issue-tp4028748p4028749.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Weaving options in 3.0.0
Is that option required if I am going to use SPIFly (which finds and registers services, but get weird errors trying to get bundles to be able to use them)? Is it something I would use instead of SPIFly? I guess I'm not sure when I would know when I'd want to turn on that feature. I'll take a look at the specs and see if anything regarding weaving hooks gives me enough of an idea. I'd really like to solve this stupid java ServiceLoader issue I'm having. -- View this message in context: http://karaf.922171.n3.nabble.com/Weaving-options-in-3-0-0-tp4028727p4028737.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Really confused regarding loading SPI resources
Right, I'm running it on Karaf 3.0.0-SNAPSHOT (despite my weird problems with certain keys not doing what they are expected), so I should be fine there. I looked at the alreadt wrapped lucene bundles and while the version I need is there, I've noticed previously that the conponents I need are mostly all wrapped. The problem, is that it doesn't look like there are any OSGi instruction in those bundles either which register the services under META-INF/services into the registry. Otherwise, other bundles can't access it. That's the same place I'm at right now. It's strange, the full version of karaf 3.0.0 is 3 meg small than what appears to be the normal distribution. I looked at the features files for some of the 3.0.0-SNAPSHOT to see if the aries spifly library was included, but all I saw were a lot of references to the 0.3 version of Aries which was confusing. I'm guess I was looking at old stuff somehow. Anyhow, I'm thinking maybe I need to deploy aries spi fly myself to get it to work. Right now it looks like some geronimo SPI classes class them, but I don't think does anything with them. Thanks, Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Really-confused-regarding-loading-SPI-resources-tp4028643p4028652.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Tabbing for command list in 3.0.0-SNAPSHOT
No pressure :) I'm just curious if it's my environment, like maybe coincidentally a jvm bug or something out of the direct scope of the karaf project. I'm also concerned that if it is a full on bug, I want to make sure it doesn't somehow slip unnoticed into a released Karaf version since that is more work for you guys and I'd like to prevent it. Plus I'm dying to play with new snapshots since the OSGi version was upgraded. :) Other than this weird console issue, and a random hiccup here and there, I have to say I've been impressed at how well the Karaf 3.0 snapshots have been running and can't wait until it's ready for release. Keep up the great work! Btw, I'll download the new snapshot on a linux laptop I have and see if the problem persists. I'll let you know since if it does work, it might be my specific environment and I'll need to figure out a good place to start debugging. Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Tabbing-for-command-list-in-3-0-0-SNAPSHOT-tp4028517p4028638.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Tabbing for command list in 3.0.0-SNAPSHOT
Ok, now I'm confused. I started up the latest Karaf 3.0.0 SNAPSHOT on Linux Mint 14 with OpenJDK 1.7, the tab key and other keys worked without issue. I think rebooted into my Windows 8 OS, did the same thing and the tab completion worked. Now, I downloaded the SNAPSHOT onto my corporate windows 7 laptop, which works fine with RC1, but it still does not work correctly with the latest snapshot. :( I wonder what the heck the difference is. Looks like I'm on jdk 1.7.0_12 32 bit and that hasn't changed since the RC I'm sure. Any ideas? Could comeone point me to where the eventlistener for the tab or other keys is created? I looked around a little but didn't see it. There much be something strange going on... -- View this message in context: http://karaf.922171.n3.nabble.com/Tabbing-for-command-list-in-3-0-0-SNAPSHOT-tp4028517p4028639.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: Tabbing for command list in 3.0.0-SNAPSHOT
I opened up a ticket for this yesterday. Has noone else seen this problem? It happens from the stock standard distribution. It would be weird if I was the only one when RC1 was working fine... maybe I'll have to do some investigating. Thanks! Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/Tabbing-for-command-list-in-3-0-0-SNAPSHOT-tp4028517p4028605.html Sent from the Karaf - User mailing list archive at Nabble.com.
Re: How can I obtain a file as a resource from within a Karaf bundle
I ran into this problem a few months back and after a lot of research, I discovered that there is a bundle specific way to load resources. The example below will allow you to load multiple xsd files from the root of the bundle if I remember correctly. Anyhow, look at the javadocs related to findEntries and related methods which will help you do what you need. Ryan -- View this message in context: http://karaf.922171.n3.nabble.com/How-can-I-obtain-a-file-as-a-resource-from-within-a-Karaf-bundle-tp4028549p4028563.html Sent from the Karaf - User mailing list archive at Nabble.com.
