On Tue, Jun 11, 2019 at 4:53 PM Giles <giles...@gmail.com> wrote:
> Unfortunately, I appear to have gone from the frying pan to the fire. I
> ran one deploy that hadn't worked previously and it did work, and I
> immediately responded with an "it's working" message. And that single
> deploy still works. But multiple others (all based on the same code, with
> only varying parameters) don't, all failing with this:
>
> Started by user Giles
> java.lang.NoSuchMethodError:
> org.eclipse.jgit.lib.Repository.getRef(Ljava/lang/String;)Lorg/eclipse/jgit/lib/Ref;
> at
> jenkins.plugins.git.GitSCMFileSystem$1.invoke(GitSCMFileSystem.java:117)
>
> That may indicate that something has installed a newer version of JGit
than JGit 4.5 which is bundled with git client plugin 2.7.x. JGit 4.5 and
prior versions provided the getRef(String) API. Newer JGit versions (like
4.9, 5.1, 5.2, and 5.3) have deprecated or stopped delivering the
getRef(String) API.
Can you confirm that the git client plugin version you're running is 2.7.x
and not one of the pre-release 3.0 versions? If it is not, then you likely
need to install the most recent 2.7 version.
Can you confirm that the git plugin version you're running is 3.9.x and not
one of the pre-release 4.0 versions? I expect that is the case since
pre-release versions of git plugin 4.0.0 generally do not refer to
getRef(String).
That message is unrelated to the version of command line git installed on
the master. That message is related to an internal mismatch between the
JGit API that is available in the running Jenkins process and the JGit API
that the git plugin requires. The git plugin is expecting to find JGit 4.5
in the Jenkins process. It does not seem to be there in this case.
This at least is a known error: it appears to happen when there's a newer
> version of git, or a misalignment of the git version and the git plugin. I
> had initially installed the very new version of Git for Windows 2.22.0. I
> downgraded that to 2.17.1.2, and eventually went to 2.21.0 - all resulting
> in the same failure. Any thoughts? It certainly doesn't seem to be the
> Git version, and the one working pipeline muddies the water further.
>
> On Tuesday, 11 June 2019 17:12:11 UTC-4, Mark Waite wrote:
>>
>> I'm glad that helped. In the past, we've seen cases where a new version
>> of ssh surprises the credential integration that is used by the git
>> plugin. I don't think your case is related to that, but I did see that Git
>> for Windows 2.22.0 now includes OpenSSH 8.0. I need to add that version
>> into my test environment. I'm not aware of any breakage with OpenSSH 8.0,
>> but we've been surprised in the past.
>>
>> On Tue, Jun 11, 2019 at 3:09 PM Giles <gile...@gmail.com> wrote:
>>
>>> Nailed it in one.
>>>
>>> I was sure it was a Jenkins issue ... but I'd upgraded Cygwin at the
>>> same time, and that probably included their version of git. Which yes, is
>>> my default git. I changed "Manage Jenkins" -> "Global Tool Configuration"
>>> -> git to the just-installed "Git for Windows", and ... fixed. I may have
>>> to make some changes in the Jenkinsfiles themselves, but it seems clear now
>>> that the latest Cygwin git broke something, and that Git for Windows fixed
>>> it ...
>>>
>>> This was crisis-level breakage for my department - I can't thank you
>>> enough.
>>>
>>> On Tuesday, 11 June 2019 16:36:38 UTC-4, Mark Waite wrote:
>>>>
>>>> Are you quite sure that your configuration did not change in about the
>>>> same time as that upgrade?
>>>>
>>>> The log file indicates that you're running command line git as provided
>>>> by Cygwin. Is that intentional? Has it always been that way?
>>>>
>>>> I don't test the Jenkins git plugin with Cygwin. I test with Git for
>>>> Windows. I'm glad cygwin git works, but don't intend to apply effort to
>>>> make it work.
>>>>
>>>> I don't see anything suspicious or troubling in the list of updates.
>>>> As far as I can tell, none of them should have changed that behavior of the
>>>> git plugin.
>>>>
>>>> On Tue, Jun 11, 2019 at 1:54 PM Giles <gile...@gmail.com> wrote:
>>>>
>>>>> I'm running Jenkins 2.164.3 on a Windows server. It's been running
>>>>> well for several months. Every Friday evening I do all the Jenkins plugin
>>>>> updates. After this past Friday's updates (details below), all our jobs
>>>>> are broken - apparently because the GitSCM checkout is broken. Our
>>>>> Jenkinsfiles are all stored in a git repo, so Jenkins attempts to pull a
>>>>> new version before running the job. This now fails. I created a test
>>>>> Jenkinsfile (editing directly in Jenkins rather than in the git repo) with
>>>>> this code in it:
>>>>>
>>>>> def scmVars = checkout(
>>>>> [
>>>>> $class: 'GitSCM',
>>>>> branches: [[name: '*/' + branch ]],
>>>>> doGenerateSubmoduleConfigurations: false,
>>>>> extensions: [],
>>>>> submoduleCfg: [],
>>>>> userRemoteConfigs: [
>>>>> [
>>>>> credentialsId: jenkinsCred,
>>>>> url: jenkinsRepo,
>>>>> ]
>>>>> ]
>>>>> ]
>>>>> );
>>>>>
>>>>> (Assume the "jenkinsCred" and "jenkinsRepo" values are correct.) The
>>>>> failure looks like this:
>>>>>
>>>>> [Pipeline] {[Pipeline] checkoutusing credential
>>>>> 29d83033-ebf6-4811-9c45-b0aadec775c2
>>>>> > C:\cygwin64\bin\git.exe rev-parse --is-inside-work-tree # timeout=10
>>>>> Fetching changes from the remote Git repository
>>>>> > C:\cygwin64\bin\git.exe config remote.origin.url
>>>>> g...@github.com:*/*.git # timeout=10
>>>>> Fetching upstream changes from g...@github.com:*/*.git
>>>>> > C:\cygwin64\bin\git.exe --version # timeout=10
>>>>> using GIT_SSH to set credentials Read-only access to the "Jenkins"
>>>>> repository at github.com/*.
>>>>> > C:\cygwin64\bin\git.exe fetch --tags --force --progress
>>>>> g...@github.com:*/*.git +refs/heads/*:refs/remotes/origin/* # timeout=10
>>>>> ERROR: Error fetching remote repo 'origin'
>>>>> hudson.plugins.git.GitException: Failed to fetch from
>>>>> g...@github.com:*/*.git
>>>>> at hudson.plugins.git.GitSCM.fetchFrom(GitSCM.java:894)
>>>>> at hudson.plugins.git.GitSCM.retrieveChanges(GitSCM.java:1161)
>>>>> at hudson.plugins.git.GitSCM.checkout(GitSCM.java:1192)
>>>>> at
>>>>> org.jenkinsci.plugins.workflow.steps.scm.SCMStep.checkout(SCMStep.java:120)
>>>>> at
>>>>> org.jenkinsci.plugins.workflow.steps.scm.SCMStep$StepExecutionImpl.run(SCMStep.java:90)
>>>>> at
>>>>> org.jenkinsci.plugins.workflow.steps.scm.SCMStep$StepExecutionImpl.run(SCMStep.java:77)
>>>>> at
>>>>> org.jenkinsci.plugins.workflow.steps.SynchronousNonBlockingStepExecution.lambda$start$0(SynchronousNonBlockingStepExecution.java:47)
>>>>> at java.util.concurrent.Executors$RunnableAdapter.call(Unknown Source)
>>>>> at java.util.concurrent.FutureTask.run(Unknown Source)
>>>>> at java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)
>>>>> at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
>>>>> at java.lang.Thread.run(Unknown Source)
>>>>> Caused by: hudson.plugins.git.GitException: Command
>>>>> "C:\cygwin64\bin\git.exe fetch --tags --force --progress
>>>>> g...@github.com:*/*.git +refs/heads/*:refs/remotes/origin/*" returned
>>>>> status code 128:
>>>>> stdout:
>>>>> stderr: error: cannot run
>>>>> D:\Jenkins\workspace\DEBUG_07_git_Checkout@tmp\jenkins-gitclient-ssh1436722960383118000.bat:
>>>>> No such file or directory
>>>>> fatal: unable to fork
>>>>>
>>>>> at
>>>>> org.jenkinsci.plugins.gitclient.CliGitAPIImpl.launchCommandIn(CliGitAPIImpl.java:2318)
>>>>> at
>>>>> org.jenkinsci.plugins.gitclient.CliGitAPIImpl.launchCommandWithCredentials(CliGitAPIImpl.java:1905)
>>>>> at
>>>>> org.jenkinsci.plugins.gitclient.CliGitAPIImpl.access$400(CliGitAPIImpl.java:81)
>>>>> at
>>>>> org.jenkinsci.plugins.gitclient.CliGitAPIImpl$1.execute(CliGitAPIImpl.java:488)
>>>>> at hudson.plugins.git.GitSCM.fetchFrom(GitSCM.java:892)
>>>>> ... 11 more[Pipeline] }[Pipeline] // node[Pipeline] End of
>>>>> PipelineERROR: Error fetching remote repo 'origin'
>>>>> Finished: FAILURE
>>>>>
>>>>> This is essentially identical to the failure we're seeing for all jobs.
>>>>> I assume it's normal behaviour that GitSCM generates and
>>>>> tries to run a random-named bat file? It's unclear to me if the bat file
>>>>> isn't created or is unreadable, although Jenkins is running
>>>>> as SYSTEM and has generally been able to read and write everything.
>>>>>
>>>>> Does anyone have any thoughts on what may have gone wrong?
>>>>>
>>>>> -----
>>>>>
>>>>>
>>>>> The updates made Friday night:
>>>>>
>>>>> >>> Artifactory 3.2.2 -> 3.2.4 [ security warning
>>>>> ]
>>>>> >>> Pipeline: API 2.34 ->
>>>>> 2.35
>>>>> >>> Pipeline: Basic Steps 2.16 ->
>>>>> 2.18
>>>>> >>> Pipeline: Declarative 1.3.8 ->
>>>>> 1.3.9
>>>>> >>> Pipeline: Declarative Extension Points API 1.3.8 ->
>>>>> 1.3.9
>>>>> >>> Pipeline: Groovy 2.69 ->
>>>>> 2.70
>>>>> >>> Pipeline: Model API 1.3.8 ->
>>>>> 1.3.9
>>>>> >>> Pipeline: Nodes and Processes 2.30 ->
>>>>> 2.31
>>>>> >>> Pipeline: SCM Step 2.7 ->
>>>>> 2.9
>>>>> >>> Pipeline: Stage Tags Metadata 1.3.8 ->
>>>>> 1.3.9
>>>>> >>> Pipeline: Step API 2.19 ->
>>>>> 2.20
>>>>> >>> Slack Notification 2.23 -> 2.24
>>>>>
>>>>> I've since rolled back all of the "Pipeline: ..." changes. The
>>>>> problem didn't change.
>>>>>
>>>>> About 40 Cygwin updates (I didn't enumerate those.)
>>>>>
>>>> --
> You received this message because you are subscribed to the Google Groups
> "Jenkins Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to jenkinsci-users+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/jenkinsci-users/18b9de6c-9c98-455e-b149-9f43b2e4011e%40googlegroups.com
> <https://groups.google.com/d/msgid/jenkinsci-users/18b9de6c-9c98-455e-b149-9f43b2e4011e%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
> For more options, visit https://groups.google.com/d/optout.
>
--
Thanks!
Mark Waite
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to jenkinsci-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/jenkinsci-users/CAO49JtEfNCbPbx85enK_y7gD2B%3D%3DcbeTE%3DVtEMLZLuSJz7MeTg%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
