Differences

This shows you the differences between two versions of the page.

Link to this comparison view

irc:1447023600 [2017/05/27 13:44] (current)
Line 1: Line 1:
 +[00:02:45] *** ChanServ sets mode: +o temporalfox
 +
 +[00:35:02] <​melvinross_>​ failed event bus messages don't provide the original message to the failure callback.. how are the rest of you implementing retry logic?
 +
 +[01:35:17] <​BadApe>​ what i can't figure out is how to get back the error code when you set message.fail(123,​string) in the return handler
 +
 +[12:13:58] <​Lecherito>​ Hey, is it possible to use docker to ship multiple urls that point to the same entry point? Like: /​maps/​{add,​delete,​edit} in container 1, /​users/​{add,​delete,​edit} in container 2, but both are called from url.com/​maps/​add,​ url.com/​users/​delete
 +
 +[12:59:47] <​AlexLehm>​ Lecherito: that should be possible with some kind of proxy (possibly outside of vertx)
 +
 +[14:29:32] <​Lecherito>​ AlexLehm: could you give me a little bit more of general info? I am not too experienced with proxies
 +
 +[14:30:36] <​Lecherito>​ Like mapping /users/add to a port like url.com:​8082/​add,​ and /maps/add to url.com:​8083/​add (?)
 +
 +[14:30:55] <​Lecherito>​ and run each container in its own port
 +
 +[14:34:16] <​AlexLehm>​ i am not so much knowledgable with docker, but I assume you have containers with different internal ip addresses, so you can probably use ipaddress:​8080/​ and ipaddress2:​8080/​
 +
 +[14:34:52] <​AlexLehm>​ i usually think that keeping the path in the proxy mappings is less confusing, so you could use ipaddress:​8080/​users/​* on the actual service as well
 +
 +[14:36:11] <​AlexLehm>​ if you have nginx or something in front of your server, this is probably a docker container as well that does the mapping to the real services
 +
 +[14:36:38] <​AlexLehm>​ like /users/* -> ipaddress:​8080/​users/​* and /maps/* -> ipaddres2:​8080/​maps/​*
 +
 +[14:39:30] <​Lecherito>​ I like that AlexLehm , I will investigate it, thanks!
 +
 +[14:39:56] <​Lecherito>​ I want to have one resource per container so I can deploy only one as needed without having to worry if i break something else
 +
 +[15:57:59] <​prathamesh_>​ hi
 +
 +[15:59:43] <​prathamesh_>​ i am running vert.x on 12 instance on 24 Core CPU machine ​ with 6 request per seconds , mostly my cpu uses is always 1% but  some times my vert.x takes 70% -80%  usage
 +
 +[15:59:52] <​prathamesh_>​ for 2-3 min
 +
 +[16:00:17] <​prathamesh_>​ ?help
 +
 +[16:00:21] <​prathamesh_>​ ?help
 +
 +[16:01:30] <​Sticky_>​ prathamesh_:​ is that 80% as in 80% of 1 core, or 80% of 24 cores?
 +
 +[16:01:48] <​Sticky_>​ and what tool tells you that
 +
 +[16:02:46] <​prathamesh_>​ all 12 cores
 +
 +[16:03:14] <​prathamesh_>​ on TOP command i can see all 12 cre 100% bust , 0% ideal
 +
 +[16:03:30] <​prathamesh_>​ i am using new relic for monitoring
 +
 +[16:04:01] <​Sticky_>​ all 12 go up together?
 +
 +[16:05:21] <​Sticky_>​ my first guess would be you get request storm that causes a huge load, alternatively it could be a garbage collection, try turning on GC logging
 +
 +[16:05:23] <​prathamesh_>​ yes all they are in always changing ​ but all between 90 % to 100%  after couple of min it automatically comes to 1%
 +
 +[16:06:42] <​Sticky_>​ and try to get a "​vmstat -n 1" during a period when this is happening and give us a paste of the output during the period. But yeah, also try turning on GC logging to rule that out
 +
 +[16:07:21] <​prathamesh_>​ ok thanx
 +
 +[16:08:12] <​Sticky_>​ if it does last a few min it is very unlikely to be GC, but you never know
 +
 +[16:09:30] <​prathamesh_>​ my same JAR file was serving 200 request per seconds , but suddanly problems started , so i reduced my traffic but still its taking CPU , ok i will check what u suggested
 +
 +[16:46:45] <​s0va>​ i'm using normal junit with junitparams and assert4j