On Fri, Mar 4, 2016 at 8:29 AM, Alfredo Deza <adeza@xxxxxxxxxx> wrote: > On Fri, Mar 4, 2016 at 2:56 AM, Martin Palma <martin@xxxxxxxx> wrote: >> For me this seems hat the java version 1.6 is already installed and we >> have to set the default java version. Here [0] is a PR which does that >> for ubuntu precise. I tested this in a simple vagrant vm and worked >> there. Could you verify that on your infrastructure this fixes the >> issue. >> >> Best, >> Martin >> >> >> [0] https://github.com/ceph/ceph-build/pull/333 >> >> (Reposting to the mailing list since my first replay got rejected >> because of formatting.) > > Thank you so much Martin! The build for Precise is already working and > we should be getting Precise binaries today. > > Once that completes I need to get those binaries signed and pushed out > to the Hammer repositories. > > I will send out an email when that happens. Both Precise and CentOS 6 binaries are up and available. Thanks again Martin for the PR! Please do let me know if anyone finds something out of place. > >> >> On Thu, Mar 3, 2016 at 10:52 PM, Alfredo Deza <adeza@xxxxxxxxxx> wrote: >>> This is as far as I got, something is odd here: >>> >>> admin@precise-test-adeza:~$ sudo apt-get install openjdk-7-jre >>> Reading package lists... Done >>> Building dependency tree >>> Reading state information... Done >>> openjdk-7-jre is already the newest version. >>> 0 upgraded, 0 newly installed, 0 to remove and 67 not upgraded. >>> admin@precise-test-adeza:~$ java -version >>> java version "1.6.0_38" >>> OpenJDK Runtime Environment (IcedTea6 1.13.10) (6b38-1.13.10-0ubuntu0.12.04.1) >>> OpenJDK 64-Bit Server VM (build 23.25-b01, mixed mode) >>> >>> So probably the wrong package? Again, I think the solution is to apply >>> this to a precise box and find out what is the missing link :) >>> >>> On Thu, Mar 3, 2016 at 4:46 PM, Alfredo Deza <adeza@xxxxxxxxxx> wrote: >>>> On Thu, Mar 3, 2016 at 3:47 PM, Martin Palma <martin@xxxxxxxx> wrote: >>>>> Hi Alfredo, >>>>> >>>>> What Java version is needed? From the "slave" playbook I see that both (yum >>>>> and apt based systems) get the same java version (7) installed, so it seems >>>>> strange that the Precise build fail. >>>>> >>>> >>>> For some reason it seems that even though it was installed the Java >>>> version reported doesn't comply: >>>> >>>> [03/03/16 19:07:20] [SSH] Checking java version of java >>>> [03/03/16 19:07:20] [SSH] java -version returned 1.6.0_38. >>>> [03/03/16 19:07:20] [SSH] Starting sftp client. >>>> [03/03/16 19:07:20] [SSH] Copying latest slave.jar... >>>> [03/03/16 19:07:20] [SSH] Copied 506,667 bytes. >>>> Expanded the channel window size to 4MB >>>> [03/03/16 19:07:20] [SSH] Starting slave process: cd >>>> "/home/jenkins-build/build" && java -jar slave.jar >>>> >>>> The full logs are here: http://dpaste.com/3VX89V8 >>>> >>>> I *believe* that something in the playbook is preventing it to get to >>>> install the java package correctly, hence the problem. >>>> >>>> This will require running the playbook against a precise box and >>>> seeing where it fails/stops >>>> >>>> >>>> >>>>> Best, >>>>> Martin >>>>> >>>>> On Thu, Mar 3, 2016 at 8:41 PM, Alfredo Deza <adeza@xxxxxxxxxx> wrote: >>>>>> >>>>>> Hi All, >>>>>> >>>>>> Some good news, we were able to go back and get CentOS 6 binaries >>>>>> built for the 0.94.6 release. >>>>>> >>>>>> They aren't published yet (they remain unsigned, queued up in our >>>>>> release process). >>>>>> >>>>>> However, we are at an impasse with the Precise build. We can't get >>>>>> Precise nodes to build because the distro has a very old version of >>>>>> the JVM which causes issues when connecting to Jenkins as a node. >>>>>> >>>>>> The exact error when attempting a connection is: >>>>>> >>>>>> Caused by: java.lang.UnsupportedClassVersionError: >>>>>> hudson/slaves/SlaveComputer$SlaveVersion: Unsupported major.minor 51.0 >>>>>> >>>>>> The way we get a new node for the build system up and running is by >>>>>> making an API call to an OpenStack provider that a boot time it runs >>>>>> some Ansible playbook which allows the new host to comply with our >>>>>> needs to be a workable node and register itself to the Master Jenkins >>>>>> server. >>>>>> >>>>>> What does this mean? --> I can't commit to fix this and would love to >>>>>> have some help from anyone in the community. I can review/merge a pull >>>>>> request with a patch to the playbook we use to setup slaves so that if >>>>>> a Precise machine is used it gets a newer JVM version. >>>>>> >>>>>> The Github project for our build system is: >>>>>> https://github.com/ceph/ceph-build >>>>>> >>>>>> The YAML file we use to setup new slaves is: >>>>>> https://github.com/ceph/ceph-build/blob/master/ansible/slave.yml >>>>>> >>>>>> Once a fix is in place I can go and get the process going again, >>>>>> complete the signing of the binaries and get them released into the >>>>>> Hammer repositories. >>>>>> >>>>>> Thanks >>>>>> >>>>>> >>>>>> Alfredo >>>>>> -- >>>>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in >>>>>> the body of a message to majordomo@xxxxxxxxxxxxxxx >>>>>> More majordomo info at http://vger.kernel.org/majordomo-info.html >>>>> >>>>> -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
