Re: [DISCUSS] Continuous integration
I would favour continuing to use Apache Jenkins for continuous integration but to disable the cargo tests on it, and then using a external CI instance with more firepower for the cargo tests. Colm. On Fri, Mar 15, 2013 at 10:58 AM, Francesco Chicchiriccò ilgro...@apache.org wrote: On 15/03/2013 11:14, Francesco Chicchiriccň wrote: Hi all, as you might have noticed, our Jenkins jobs [1] are quite unstable, lately more than ever. This is not due to actual test failures but (almost always) to the inability of Jenkins nodes to start Tomcat via cargo within 2 minutes (the default for cargo, 12 msecs). And this is in turn caused by the high load under which the ASF Jenkins infrastructure is working; lately more than ever. Having seen this trend, I have recently added some Syncope jobs to our company Jenkins instance [2]; this in order to have a more reliable reference to understand at first glance whether an ASF Jenkins failure is actually due to test failures or not. But this is not satisfying for different reasons: 1. it is a private - even though publicly accessible - Jenkins instance: for this reason, for example, it is not configured for sending messages to dev@syncope.apache.org (as instead ASF Jenkins does) 2. it cannot access ASF JIRA to publish test results as an issue comment (is that true?) 3. we don't have enough bandwidth to continuously publish Maven artifacts to ASF snapshot repository 4. I don't know if this is allowed by ASF policies If there is no way out of this ASF Jenkins trouble, I'd propose then to find some other free CI over the Internet (Cloudbees Jenkins [3] or Travis CI [4]). Forgot to add: Jackrabbit is already using Travis CI [5] so there shouldn't be any formal problem with that; only, we need that our repository is also mirrored to Github, as Colm requested a while ago [6]. Regards. [1] https://builds.apache.org/**view/S-Z/view/Syncope/https://builds.apache.org/view/S-Z/view/Syncope/ [2] http://jenkins.tirasa.net/ [3] http://www.cloudbees.com/**jenkins-enterprise-by-** cloudbees-overview.cbhttp://www.cloudbees.com/jenkins-enterprise-by-cloudbees-overview.cb [4] https://travis-ci.org/ [5] https://travis-ci.org/apache/**jackrabbit-oakhttps://travis-ci.org/apache/jackrabbit-oak [6] https://issues.apache.org/**jira/browse/INFRA-5837https://issues.apache.org/jira/browse/INFRA-5837 -- Francesco Chicchiriccň ASF Member, Apache Syncope PMC chair, Apache Cocoon PMC Member http://people.apache.org/~**ilgrosso/http://people.apache.org/~ilgrosso/ -- Colm O hEigeartaigh Talend Community Coder http://coders.talend.com
Re: [DISCUSS] Continuous integration
On 19/03/2013 11:38, Colm O hEigeartaigh wrote: I would favour continuing to use Apache Jenkins for continuous integration but to disable the cargo tests on it, and then using a external CI instance with more firepower for the cargo tests. +1 looks reasonable to me. My choice goes to Travis CI, but we need that the mirroring to github is enabled: do you have any news for this? Regards. On Fri, Mar 15, 2013 at 10:58 AM, Francesco Chicchiriccò ilgro...@apache.org wrote: On 15/03/2013 11:14, Francesco Chicchiriccň wrote: Hi all, as you might have noticed, our Jenkins jobs [1] are quite unstable, lately more than ever. This is not due to actual test failures but (almost always) to the inability of Jenkins nodes to start Tomcat via cargo within 2 minutes (the default for cargo, 12 msecs). And this is in turn caused by the high load under which the ASF Jenkins infrastructure is working; lately more than ever. Having seen this trend, I have recently added some Syncope jobs to our company Jenkins instance [2]; this in order to have a more reliable reference to understand at first glance whether an ASF Jenkins failure is actually due to test failures or not. But this is not satisfying for different reasons: 1. it is a private - even though publicly accessible - Jenkins instance: for this reason, for example, it is not configured for sending messages to dev@syncope.apache.org (as instead ASF Jenkins does) 2. it cannot access ASF JIRA to publish test results as an issue comment (is that true?) 3. we don't have enough bandwidth to continuously publish Maven artifacts to ASF snapshot repository 4. I don't know if this is allowed by ASF policies If there is no way out of this ASF Jenkins trouble, I'd propose then to find some other free CI over the Internet (Cloudbees Jenkins [3] or Travis CI [4]). Forgot to add: Jackrabbit is already using Travis CI [5] so there shouldn't be any formal problem with that; only, we need that our repository is also mirrored to Github, as Colm requested a while ago [6]. Regards. [1] https://builds.apache.org/**view/S-Z/view/Syncope/https://builds.apache.org/view/S-Z/view/Syncope/ [2] http://jenkins.tirasa.net/ [3] http://www.cloudbees.com/**jenkins-enterprise-by-** cloudbees-overview.cbhttp://www.cloudbees.com/jenkins-enterprise-by-cloudbees-overview.cb [4] https://travis-ci.org/ [5] https://travis-ci.org/apache/**jackrabbit-oakhttps://travis-ci.org/apache/jackrabbit-oak [6] https://issues.apache.org/**jira/browse/INFRA-5837https://issues.apache.org/jira/browse/INFRA-5837 -- Francesco Chicchiriccò ASF Member, Apache Syncope PMC chair, Apache Cocoon PMC Member http://people.apache.org/~ilgrosso/
Re: [DISCUSS] Continuous integration
No, not yet. Still waiting for some feedback from INFRA here: https://issues.apache.org/jira/browse/INFRA-5837 Colm. On Tue, Mar 19, 2013 at 10:40 AM, Francesco Chicchiriccò ilgro...@apache.org wrote: On 19/03/2013 11:38, Colm O hEigeartaigh wrote: I would favour continuing to use Apache Jenkins for continuous integration but to disable the cargo tests on it, and then using a external CI instance with more firepower for the cargo tests. +1 looks reasonable to me. My choice goes to Travis CI, but we need that the mirroring to github is enabled: do you have any news for this? Regards. On Fri, Mar 15, 2013 at 10:58 AM, Francesco Chicchiriccò ilgro...@apache.org wrote: On 15/03/2013 11:14, Francesco Chicchiriccň wrote: Hi all, as you might have noticed, our Jenkins jobs [1] are quite unstable, lately more than ever. This is not due to actual test failures but (almost always) to the inability of Jenkins nodes to start Tomcat via cargo within 2 minutes (the default for cargo, 12 msecs). And this is in turn caused by the high load under which the ASF Jenkins infrastructure is working; lately more than ever. Having seen this trend, I have recently added some Syncope jobs to our company Jenkins instance [2]; this in order to have a more reliable reference to understand at first glance whether an ASF Jenkins failure is actually due to test failures or not. But this is not satisfying for different reasons: 1. it is a private - even though publicly accessible - Jenkins instance: for this reason, for example, it is not configured for sending messages to dev@syncope.apache.org (as instead ASF Jenkins does) 2. it cannot access ASF JIRA to publish test results as an issue comment (is that true?) 3. we don't have enough bandwidth to continuously publish Maven artifacts to ASF snapshot repository 4. I don't know if this is allowed by ASF policies If there is no way out of this ASF Jenkins trouble, I'd propose then to find some other free CI over the Internet (Cloudbees Jenkins [3] or Travis CI [4]). Forgot to add: Jackrabbit is already using Travis CI [5] so there shouldn't be any formal problem with that; only, we need that our repository is also mirrored to Github, as Colm requested a while ago [6]. Regards. [1] https://builds.apache.org/view/S-Z/view/Syncope/https://builds.apache.org/**view/S-Z/view/Syncope/ https:/**/builds.apache.org/view/S-Z/**view/Syncope/https://builds.apache.org/view/S-Z/view/Syncope/ [2] http://jenkins.tirasa.net/ [3] http://www.cloudbees.com/jenkins-enterprise-by-**http://www.cloudbees.com/**jenkins-enterprise-by-** cloudbees-overview.cbhttp://**www.cloudbees.com/jenkins-** enterprise-by-cloudbees-**overview.cbhttp://www.cloudbees.com/jenkins-enterprise-by-cloudbees-overview.cb [4] https://travis-ci.org/ [5] https://travis-ci.org/apache/jackrabbit-oakhttps://travis-ci.org/apache/**jackrabbit-oak https://**travis-ci.org/apache/**jackrabbit-oakhttps://travis-ci.org/apache/jackrabbit-oak [6] https://issues.apache.org/jira/browse/INFRA-5837https://issues.apache.org/**jira/browse/INFRA-5837 https:/**/issues.apache.org/jira/**browse/INFRA-5837https://issues.apache.org/jira/browse/INFRA-5837 -- Francesco Chicchiriccò ASF Member, Apache Syncope PMC chair, Apache Cocoon PMC Member http://people.apache.org/~**ilgrosso/http://people.apache.org/~ilgrosso/ -- Colm O hEigeartaigh Talend Community Coder http://coders.talend.com
Re: [DISCUSS] Continuous integration
On 17/03/2013 10:42, Andrei Shakirin wrote: Hi Francesco, Do you see any ways to resolve the problem through: a) increase default 2 minutes timeout This is feasible but I don't know if such timeout makes sense when 10 minutes (this is the kind of performance available of ASF Jenkins nowadays...) b) optimize cargo Syncope startup (not sure is it really possible) Don't have idea of what to do in this respect... Regards. -Original Message- From: Francesco Chicchiriccò [mailto:ilgro...@apache.org] Sent: Freitag, 15. März 2013 11:15 To: dev@syncope.apache.org Subject: [DISCUSS] Continuous integration Hi all, as you might have noticed, our Jenkins jobs [1] are quite unstable, lately more than ever. This is not due to actual test failures but (almost always) to the inability of Jenkins nodes to start Tomcat via cargo within 2 minutes (the default for cargo, 12 msecs). And this is in turn caused by the high load under which the ASF Jenkins infrastructure is working; lately more than ever. Having seen this trend, I have recently added some Syncope jobs to our company Jenkins instance [2]; this in order to have a more reliable reference to understand at first glance whether an ASF Jenkins failure is actually due to test failures or not. But this is not satisfying for different reasons: 1. it is a private - even though publicly accessible - Jenkins instance: for this reason, for example, it is not configured for sending messages to dev@syncope.apache.org (as instead ASF Jenkins does) 2. it cannot access ASF JIRA to publish test results as an issue comment (is that true?) 3. we don't have enough bandwidth to continuously publish Maven artifacts to ASF snapshot repository 4. I don't know if this is allowed by ASF policies If there is no way out of this ASF Jenkins trouble, I'd propose then to find some other free CI over the Internet (Cloudbees Jenkins [3] or Travis CI [4]). WDYT? [1] https://builds.apache.org/view/S-Z/view/Syncope/ [2] http://jenkins.tirasa.net/ [3] http://www.cloudbees.com/jenkins-enterprise-by-cloudbees- overview.cb [4] https://travis-ci.org/ -- Francesco Chicchiriccò ASF Member, Apache Syncope PMC chair, Apache Cocoon PMC Member http://people.apache.org/~ilgrosso/
Re: [DISCUSS] Continuous integration
Il 15/03/2013 11:14, Francesco Chicchiriccò ha scritto: Hi all, as you might have noticed, our Jenkins jobs [1] are quite unstable, lately more than ever. This is not due to actual test failures but (almost always) to the inability of Jenkins nodes to start Tomcat via cargo within 2 minutes (the default for cargo, 12 msecs). And this is in turn caused by the high load under which the ASF Jenkins infrastructure is working; lately more than ever. Having seen this trend, I have recently added some Syncope jobs to our company Jenkins instance [2]; this in order to have a more reliable reference to understand at first glance whether an ASF Jenkins failure is actually due to test failures or not. But this is not satisfying for different reasons: 1. it is a private - even though publicly accessible - Jenkins instance: for this reason, for example, it is not configured for sending messages to dev@syncope.apache.org (as instead ASF Jenkins does) 2. it cannot access ASF JIRA to publish test results as an issue comment (is that true?) 3. we don't have enough bandwidth to continuously publish Maven artifacts to ASF snapshot repository 4. I don't know if this is allowed by ASF policies If there is no way out of this ASF Jenkins trouble, I'd propose then to find some other free CI over the Internet (Cloudbees Jenkins [3] or Travis CI [4]). WDYT? ASF policies permitting: +1 [1] https://builds.apache.org/view/S-Z/view/Syncope/ [2] http://jenkins.tirasa.net/ [3] http://www.cloudbees.com/jenkins-enterprise-by-cloudbees-overview.cb [4] https://travis-ci.org/ -- Massimiliano Perrone Tel +39 393 9121310 Tirasa S.r.l. Viale D'Annunzio 267 - 65127 Pescara Tel +39 0859116307 / FAX +39 085973 http://www.tirasa.net Apache Syncope PMC Member http://people.apache.org/~massi/ L'apprendere molte cose non insegna l'intelligenza (Eraclito)
Re: [DISCUSS] Continuous integration
On 15/03/2013 11:14, Francesco Chicchiriccò wrote: Hi all, as you might have noticed, our Jenkins jobs [1] are quite unstable, lately more than ever. This is not due to actual test failures but (almost always) to the inability of Jenkins nodes to start Tomcat via cargo within 2 minutes (the default for cargo, 12 msecs). And this is in turn caused by the high load under which the ASF Jenkins infrastructure is working; lately more than ever. Having seen this trend, I have recently added some Syncope jobs to our company Jenkins instance [2]; this in order to have a more reliable reference to understand at first glance whether an ASF Jenkins failure is actually due to test failures or not. But this is not satisfying for different reasons: 1. it is a private - even though publicly accessible - Jenkins instance: for this reason, for example, it is not configured for sending messages to dev@syncope.apache.org (as instead ASF Jenkins does) 2. it cannot access ASF JIRA to publish test results as an issue comment (is that true?) 3. we don't have enough bandwidth to continuously publish Maven artifacts to ASF snapshot repository 4. I don't know if this is allowed by ASF policies If there is no way out of this ASF Jenkins trouble, I'd propose then to find some other free CI over the Internet (Cloudbees Jenkins [3] or Travis CI [4]). Forgot to add: Jackrabbit is already using Travis CI [5] so there shouldn't be any formal problem with that; only, we need that our repository is also mirrored to Github, as Colm requested a while ago [6]. Regards. [1] https://builds.apache.org/view/S-Z/view/Syncope/ [2] http://jenkins.tirasa.net/ [3] http://www.cloudbees.com/jenkins-enterprise-by-cloudbees-overview.cb [4] https://travis-ci.org/ [5] https://travis-ci.org/apache/jackrabbit-oak [6] https://issues.apache.org/jira/browse/INFRA-5837 -- Francesco Chicchiriccò ASF Member, Apache Syncope PMC chair, Apache Cocoon PMC Member http://people.apache.org/~ilgrosso/