Re: [Artemis] Error message in log file for cluster on v1.3
I can check if by directly setting SSL example with static connectors. This error message is appearing only when cluster is setup with TLS. -- View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-Error-message-in-log-file-for-cluster-on-v1-3-tp4713397p4728619.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Artemis - Wild card - messages received but not acknowledged
A reproducible test-case is always the best way to get help on issues like this. Artemis ships with lots of examples. It's typically fairly trivial to modify one of those examples to reproduce straight-forward use-cases. You can share the modified example via your GitHub repo. Could you work up something like this? At the very least could you provide specific steps/code/configuration to reproduce even if it isn't automated? Justin On Mon, Jul 17, 2017 at 3:37 PM, praneethgwrote: > I have a consumer running and listening to aes.q.#, > Messages sent to aes.q.1 , aes.q.2 are consumed by consumer but not > acknowledged. Message still remains in queue. > But if i listen to those specific queues aes.q.1, then message is received > and acknowledged. > > any one using wild card on the consumer side for artemis? > > > > -- > View this message in context: http://activemq.2283324.n4. > nabble.com/Artemis-Wild-card-messages-received-but-not- > acknowledged-tp4728617.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. >
Artemis - Wild card - messages received but not acknowledged
I have a consumer running and listening to aes.q.#, Messages sent to aes.q.1 , aes.q.2 are consumed by consumer but not acknowledged. Message still remains in queue. But if i listen to those specific queues aes.q.1, then message is received and acknowledged. any one using wild card on the consumer side for artemis? -- View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-Wild-card-messages-received-but-not-acknowledged-tp4728617.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: [Artemis] Error message in log file for cluster on v1.3
Is all that really necessary to reproduce the problem? Part of the exercise in reproducing the problem is simplifying the user-case and configuration down to only what is absolutely necessary to duplicate the issue. Justin On Mon, Jul 17, 2017 at 1:56 PM, abhijithwrote: > Yep will do that. Will it be fine If I give a docker compose file based on > https://github.com/vromero/activemq-artemis-docker? > > I need to bring up couple of nginx load balancers with ssl offloading as > well. I will set it up with self signed certs > > > > > -- > View this message in context: http://activemq.2283324.n4. > nabble.com/Artemis-Error-message-in-log-file-for-cluster-on-v1-3- > tp4713397p4728611.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. >
Re: [Artemis] Error message in log file for cluster on v1.3
Yep will do that. Will it be fine If I give a docker compose file based on https://github.com/vromero/activemq-artemis-docker? I need to bring up couple of nginx load balancers with ssl offloading as well. I will set it up with self signed certs -- View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-Error-message-in-log-file-for-cluster-on-v1-3-tp4713397p4728611.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: [Artemis] Error message in log file for cluster on v1.3
Artemis ships with a statically configured HA example called "replicated-failback-static" that's working fine in the 2.1.0 release. Perhaps you could provide a reproducible test-case based on that example. Justin On Mon, Jul 17, 2017 at 11:31 AM, Justin Bertramwrote: > Do you have a reproducible test-case or anything more you could share with > regards to configuration? Right now we just have some log messages and > high level explanation of the configuration. That's not really enough to > diagnose the problem. > > > Justin > > On Mon, Jul 17, 2017 at 10:46 AM, abhijith > wrote: > >> Resurrecting this thread as we are still facing this issue in v2.1. Any >> suggestions on how to take it out? There is no issue is performance or >> message delivery except filling up of logs. Its very annoying and makes >> debugging almost impossible. >> >> >> >> -- >> View this message in context: http://activemq.2283324.n4.nab >> ble.com/Artemis-Error-message-in-log-file-for-cluster-on-v1- >> 3-tp4713397p4728605.html >> Sent from the ActiveMQ - User mailing list archive at Nabble.com. >> > >
Re: [Artemis] Error message in log file for cluster on v1.3
Do you have a reproducible test-case or anything more you could share with regards to configuration? Right now we just have some log messages and high level explanation of the configuration. That's not really enough to diagnose the problem. Justin On Mon, Jul 17, 2017 at 10:46 AM, abhijithwrote: > Resurrecting this thread as we are still facing this issue in v2.1. Any > suggestions on how to take it out? There is no issue is performance or > message delivery except filling up of logs. Its very annoying and makes > debugging almost impossible. > > > > -- > View this message in context: http://activemq.2283324.n4. > nabble.com/Artemis-Error-message-in-log-file-for-cluster-on-v1-3- > tp4713397p4728605.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. >
Re: [Artemis] Error message in log file for cluster on v1.3
Resurrecting this thread as we are still facing this issue in v2.1. Any suggestions on how to take it out? There is no issue is performance or message delivery except filling up of logs. Its very annoying and makes debugging almost impossible. -- View this message in context: http://activemq.2283324.n4.nabble.com/Artemis-Error-message-in-log-file-for-cluster-on-v1-3-tp4713397p4728605.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Transport scheme NOT recognized: [mqtt+nio]
You have to add the fusesource MQTT client to your dependency manager, as it seem they forgot to include it in the activemq-all.jar file Example with maven pom.xml org.fusesource.mqtt-client mqtt-client 1.12 -- View this message in context: http://activemq.2283324.n4.nabble.com/Transport-scheme-NOT-recognized-mqtt-nio-tp4724931p4728604.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: Active Durable Subscriber status automatically changing to offline Durable Subscriber [5.14.1 Most Stable Version]
Thank you so much for your response. Issue is resolved after removing that parameter. Thanks & Regards, Tejas Ramchandra Sawant Tata Consultancy Services Ltd. QBPL, Phase-2, Hinjewadi Pune, Maharashtra. cell : +91-8055946458 Mailto: tejas.sawa...@tcs.com Website:http://www.tcs.com Experience certainty.IT Services Business Solutions Consulting From: "Tim Bain [via ActiveMQ]"To: tejas13 Date: 07/17/2017 06:18 PM Subject:Re: Active Durable Subscriber status automatically changing to offline Durable Subscriber [5.14.1 Most Stable Version] A maxInactivityDuration of 0 will disable inactivity detection, whose purpose is to detect both idle TCP connections and TCP connections that died ungracefully (e.g. due to network partitions), so I believe that your addition of that parameter caused this behavior. What was the reason you set it to that value? Tim On Jul 15, 2017 1:11 AM, "tejas13" <[hidden email]> wrote: > > > When this happens with your real-world server (in the morning or the middle > of the night), are you experiencing network partitions that result in the > subscriber really truly being offline for some period of time? If so, you > completely left that out of your initial description, and it means that > your question is actually, "Why doesn't my consumer reconnect automatically > after a network partition?" If not, why are you doing that during the test > you described here? > > Network Partition is one reason. There may be another reason also > > With 5.12.3, is SUB1 seen as offline in step 8? That is, is the only > difference in behavior between 5.12.3 and 5.14.5 what you see in Step 11, > but you see identical behavior for all other steps? > > In case of 5.12.3 I found both subscriber offline when network cable > disconnected and online when connected > > > Is your consumer using a URI that uses the failover transport, to ensure it > automatically reconnects after the network connection is lost? More > generally, what URI is your client using to connect to the broker? > > Yes I am using fail over protocol > > At what Log4J level is your webapp logging for ActiveMQ-related loggers? > What did you see when you switched them to DEBUG? I find it very surprising > that you would not see any logging by the client when you pull its network > cable; I'd expect that there'd be some logging that would occur after your > TCP session timeout elapses (did you wait that long?) and the client > figures out that it's not connected to the broker, so if you're not seeing > anything (Step 6) then I'd wonder whether you have Log4J configured > correctly. > > Yes in case of 5.12.3 I get exception in application. But in 5.14.5 I am > not getting any error also. > > Is there any difference between what's in the logs (either broker or > webapp) when you use the 5.14.5 client vs. when you use the 5.12.3 client? > Also, when you said you use 5.12.3 or 5.14.5 for a given test run, are you > switching both the broker and the client JARs? > > I am using 5.14.3 jars. > > One change we did recently in activemq.xml and we faced this issue. That > change is changing open wire protocol from tcp to nio and added parameter > maxInactivityDuration > > > > > Please let me know does this is impacting solution. > > > Thanks & Regards, > Tejas Ramchandra Sawant > Tata Consultancy Services Ltd. > QBPL, Phase-2, Hinjewadi > Pune, Maharashtra. > cell : +91-8055946458 > Mailto: [hidden email] > Website:http://www.tcs.com > > > Experience certainty.IT Services > Business Solutions > Consulting > > > > -"Tim Bain [via ActiveMQ]" <[hidden email]> > wrote: - > To: tejas13 <[hidden email]> > From: "Tim Bain [via ActiveMQ]" <[hidden email]> > Date: 07/15/2017 10:30AM > Subject: Re: Active Durable Subscriber status automatically changing to > offline Durable Subscriber [5.14.1 Most Stable Version] > > When this happens with your real-world server (in the morning or the middle > of the night), are you experiencing network partitions that result in the > subscriber really truly being offline for some period of time? If so, you > completely left that out of your initial description, and it means that > your question is actually, "Why doesn't my consumer reconnect automatically > after a network partition?" If not, why are you doing that during the test > you described here? > > With 5.12.3, is SUB1 seen as offline in step 8? That is, is the only > difference in behavior between 5.12.3 and 5.14.5 what you see in Step 11, > but you see identical behavior
Re: How to SYNC messages in Cluster without using persistence on the message broker
This is ActiveMQ 5.x, not Artemis, right? What do you mean by "SYNC"? If you mean "replicate the message so a copy exists on all three brokers and can be consumed from any of them, and delete it from all three when it's consumed from one", ActiveMQ 5.x doesn't do that. Artemis, however, does support clustered operation, so you might want to consider switching to it. If you mean something else, please describe the behavior you're hoping to achieve. Tim On Jul 17, 2017 6:15 AM, "Senthil Jayakumar"wrote: > Hi There, > > I'm wondering if there is any way to sync/replicate the message without > persisting on the message broker. > > I have THREE node ODL cluster and ActiveMQ runs on all the nodes (without > knowing each other). There are few topics created and message producer on > one of the nodes and would like to check if there is any way to SYNC the > topics and the messages on other two nodes without using persistence. > > Thanks, > Senthil Kumar Jayakumar >
Re: Active Durable Subscriber status automatically changing to offline Durable Subscriber [5.14.1 Most Stable Version]
A maxInactivityDuration of 0 will disable inactivity detection, whose purpose is to detect both idle TCP connections and TCP connections that died ungracefully (e.g. due to network partitions), so I believe that your addition of that parameter caused this behavior. What was the reason you set it to that value? Tim On Jul 15, 2017 1:11 AM, "tejas13"wrote: > > > When this happens with your real-world server (in the morning or the middle > of the night), are you experiencing network partitions that result in the > subscriber really truly being offline for some period of time? If so, you > completely left that out of your initial description, and it means that > your question is actually, "Why doesn't my consumer reconnect automatically > after a network partition?" If not, why are you doing that during the test > you described here? > > Network Partition is one reason. There may be another reason also > > With 5.12.3, is SUB1 seen as offline in step 8? That is, is the only > difference in behavior between 5.12.3 and 5.14.5 what you see in Step 11, > but you see identical behavior for all other steps? > > In case of 5.12.3 I found both subscriber offline when network cable > disconnected and online when connected > > > Is your consumer using a URI that uses the failover transport, to ensure it > automatically reconnects after the network connection is lost? More > generally, what URI is your client using to connect to the broker? > > Yes I am using fail over protocol > > At what Log4J level is your webapp logging for ActiveMQ-related loggers? > What did you see when you switched them to DEBUG? I find it very surprising > that you would not see any logging by the client when you pull its network > cable; I'd expect that there'd be some logging that would occur after your > TCP session timeout elapses (did you wait that long?) and the client > figures out that it's not connected to the broker, so if you're not seeing > anything (Step 6) then I'd wonder whether you have Log4J configured > correctly. > > Yes in case of 5.12.3 I get exception in application. But in 5.14.5 I am > not getting any error also. > > Is there any difference between what's in the logs (either broker or > webapp) when you use the 5.14.5 client vs. when you use the 5.12.3 client? > Also, when you said you use 5.12.3 or 5.14.5 for a given test run, are you > switching both the broker and the client JARs? > > I am using 5.14.3 jars. > > One change we did recently in activemq.xml and we faced this issue. That > change is changing open wire protocol from tcp to nio and added parameter > maxInactivityDuration > > > > > Please let me know does this is impacting solution. > > > Thanks & Regards, > Tejas Ramchandra Sawant > Tata Consultancy Services Ltd. > QBPL, Phase-2, Hinjewadi > Pune, Maharashtra. > cell : +91-8055946458 > Mailto: tejas.sawa...@tcs.com > Website:http://www.tcs.com > > > Experience certainty.IT Services > Business Solutions > Consulting > > > > -"Tim Bain [via ActiveMQ]" > wrote: - > To: tejas13 > From: "Tim Bain [via ActiveMQ]" > Date: 07/15/2017 10:30AM > Subject: Re: Active Durable Subscriber status automatically changing to > offline Durable Subscriber [5.14.1 Most Stable Version] > > When this happens with your real-world server (in the morning or the middle > of the night), are you experiencing network partitions that result in the > subscriber really truly being offline for some period of time? If so, you > completely left that out of your initial description, and it means that > your question is actually, "Why doesn't my consumer reconnect automatically > after a network partition?" If not, why are you doing that during the test > you described here? > > With 5.12.3, is SUB1 seen as offline in step 8? That is, is the only > difference in behavior between 5.12.3 and 5.14.5 what you see in Step 11, > but you see identical behavior for all other steps? > > Is your consumer using a URI that uses the failover transport, to ensure it > automatically reconnects after the network connection is lost? More > generally, what URI is your client using to connect to the broker? > > At what Log4J level is your webapp logging for ActiveMQ-related loggers? > What did you see when you switched them to DEBUG? I find it very surprising > that you would not see any logging by the client when you pull its network > cable; I'd expect that there'd be some logging that would occur after your > TCP session timeout elapses (did you wait that long?) and the client > figures out that it's not connected to the broker, so if you're not seeing > anything (Step 6) then I'd wonder whether you have Log4J configured > correctly. > > Is there any difference between what's in the
How to SYNC messages in Cluster without using persistence on the message broker
Hi There, I'm wondering if there is any way to sync/replicate the message without persisting on the message broker. I have THREE node ODL cluster and ActiveMQ runs on all the nodes (without knowing each other). There are few topics created and message producer on one of the nodes and would like to check if there is any way to SYNC the topics and the messages on other two nodes without using persistence. Thanks, Senthil Kumar Jayakumar
Re: Active MQ - runtimeConfigurationPlugin - Set Message Priority - Not working
This is what is updated in the JIRA (FYI) - this not going to be fixed, may be: the priority support is not a good candidate for runtime modification because not only do the in-memory messages need to retained in priority memory, the store needs to retrieve and store messages in priority order. For this, the store needs a restart which is not practical. -- View this message in context: http://activemq.2283324.n4.nabble.com/Active-MQ-runtimeConfigurationPlugin-Set-Message-Priority-Not-working-tp4728264p4728590.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
client abort and can't connect server immediately
Hi, I am new to activemq artemis, I am using mqtt protocol. I use eclipse paho java client library to connect server. I find that when my program connect to server than then kill the program, (my program has no chance to call disconnect) then the client can't connect to server again, the client says: invalid client id. after for a while, maybe 1 minutes later, my program can connect again. any suggestion? Thanks advance. Shawn
Re: Active MQ - runtimeConfigurationPlugin - Set Message Priority - Not working
Thanks Tim. Created https://issues.apache.org/jira/browse/AMQ-6772 -- View this message in context: http://activemq.2283324.n4.nabble.com/Active-MQ-runtimeConfigurationPlugin-Set-Message-Priority-Not-working-tp4728264p4728584.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.