hat.
Normally you would configure the job to poll the version control
system so it can check out and build when it detects a change, and the
build script/batch would be included in the checkout, but if you want,
you set a job to build periodically and
run a windows batch file as the only build step.
ve more explicit control of the steps.
--
Les Mikesell
lesmikes...@gmail.com
--
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+unsubs
to unroll the matrix into separate jobs or
jobs with parameters and having control over when each thing happens
and which are in parallel.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
.org/display/JENKINS/Multiple+SCMs+Plugin
Yes, that should work automatically, although I'm not sure what
happens if the 2nd repo needs different credentials.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Je
eated by different users? Also, how
will it extend to svn externals if they pull in components that need
different credentials?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To
access to components referenced by externals more open
than the probably should be to accommodate access from jenkins.
Matching the credentials for the most-specific path specified by the
user seems right to me.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message
HG
> respectively, so this is a problem they face too)
'User that triggered' is an odd concept for scheduled stuff - weirdly
intertwined with the user that made a commit that affects the job
(perhaps the actual script that runs the build), and there's no
general assumption you ca
somewhere? I'm
picturing someone who works on a bunch of different projects having to
enter their credentials more than once.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscr
current set of plugins, you should be able to just
copy the hpi/jpi files out of a working jenkins's plugins directory
and drop into the new one - with some level of concern about matching
versions with the new jenkins core you are installing - but if you are
on the LTS track that doesn&
n limit, but you might hit various
operating system per-user limits like the number of open files.
Tuning them will, of course depend on your host operating system.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jen
arent, and (b) to collate all of the artifacts under
the parent job?
--
Les Mikesell
lesmikes...@gmail.com
--
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, se
ion if you feel like writing some code.
I was hoping to use one of the existing plugins for that - maybe 'Copy
To Slave' if I can get the parent to save the child build numbers so
it can get the right ones.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message beca
e, but my 2 cents is that if it matters, you are doing
something wrong or you should restrict the job to the node(s) where
you want it. Maybe there was a network glitch or something that made
jenkins think that node was unavailable.
--
Les Mikesell
lesmikes...@gmail.comn
--
You rece
the nodes with that
label. This approach will do what you want and also allow you to add
different types of nodes for different jobs. Note that nodes can have
multiple labels and jobs can require some combination of labels if you
have a complex set of capabilities and requirements, but
her job do it? I can't find a way to use the
copy artifact plugin from a build flow job.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and stop re
#x27;copy to slave' plugin, but none of my jobs run on the
master node so that doesn't make sense. (But maybe I could run the
build flow jobs there if that would make it work...).
Is it possible to do the copy within the build flow dsl itself? It
seems to handle some groovy stuff.
-
ike you should be able to run the top-level job and have that
be the place where the artifacts show up (if you want) rather than
having some other job get them.
Or are you saying I should code up the whole sequence directly in
groovy instead of using build flow at all if I need the artifacts?
--
ment, so it doesn't make sense to me for the resulting
artifacts to have to end up in some other job's space. Could the DSL
be extended to have a 'copy child artifact' function so that after the
scheduling happen you can have the results where you'd expect jenkins
to have t
he jenkins job
view containing the svn version checkout that triggered the job and
'archive artifacts' post-build step works to save any of those files
into the master archive (we usually archive the .h files along with
built libraries). Can't it copy into that space?
--
Les Mi
I don't understand how a job without a workspace or
archive coordinates with the rest of jenkins.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group an
s plus each platform's .so/.dll/etc.).
--
Les Mikesell
lesmikes...@gmail.com
--
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.
y polls. If there is no
other way to do it, should I think in terms of running one platform's
builds with the stock svn polling, then having that job kick off the
build flow job for the rest? And if I did that, can I pass the
triggering SVN_REVISION through to its child jobs so that the build
have an equivalent. I'm beginning to think that what I
really need is some boilerplate groovy code or library to do the same
operations as the build flow plugin but in an ordinary build step.
Would the 'run other jobs in parallel' feature be possible there?
--
Les Mikesell
it
> post-commit hook, then pass the incoming revision as parameter to other jobs
I'm not really clear on how matrix jobs handle this either - which is
part of the reason I'm looking for more explicit control of the child
jobs.
--
Les Mikesell
lesmikes...@gmail.com
--
You receive
e JDK versions and select them by
job. In the global jenkins configuration under JDK installations,
give each JDK a name, then in the node configurations under tool
locations, add the appropriate path to each named JDK installation.
And in the job, pick the JDK version by name.
--
Les Mikesell
or at least
the functions it offers into a library that would be available in any
groovy build step? That could match the convenience without imposing
the restrictions of a separate plugin.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed
leave it that way? You can give nodes more than one label if you are
doing something more complicated.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this gro
main slave, by saying that they have to run on the node that maches the
> value of my global property MAIN_SLAVE.
I'm not sure if there is a way to expand a value dynamically, but it
should only take 2 edits to remove a label from a node and add it to
another.
--
Les Mikesell
l
, if there is some logic behind the sequence of events you need,
maybe you could use the rest api to trigger a parameterized build with
the node being passed as a parameter - perhaps you already have some
program that understands or controls the state of things.
--
Les Mikesell
lesmikes...@gmai
ut jenkins you
might want the workspace on a different drive or to have 2 slightly
different versions checked out at once. Unless this is a problem with
Visual Source Safe I'd try that approach instead of trying to make
jenkins do it the wrong way.
--
Les Mikesell
lesmikes...@gmail
rrectly.
They aren't 'system' settings, they are interactive login settings.
You should be able to set it in the jenkins user's .bashrc file or in
the node configuration's node properties 'environment variables',
though.
--
Les Mikesell
lesmikes...@gmail.com
ve paths. Not sure how
to match that with a different VCS. Maybe just expanding the path to
an absolute location in an environment variable can do what you need.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"J
To remind, I am running jenkins on windows.
Groovy is very portable - the OS shouldn't matter.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and stop r
.
> Instead of using Tomcat, can I somehow load my .war to the Jenkins
> installation and use its web server instead?
I don't think so, but it should work the other way around. That is,
switch to the war version of jenkins and drop it under tomcat if you
want everything under one se
t
> java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
> at java.lang.Thread.run(Thread.java:662)
>
>
>
> Any ideas about this issue ?
What happens if you
.)
The built in support should take care of getting the right jar in the
right place for you - and probably doing everything else jenkins
needs...
In any case, I'd run under different users with different home
directories so the jars and other files don't clobber each other.
--
Les M
til
the whole job completes. In any case I think you have to explain why
having separate jobs to control what you want done on separate
machines is a problem before anyone could answer your philosophical
question.There's not really a limit to the number of jobs you can
configure and if
do
builds every time a chance is committed to a source repository without
complex dependencies on the order of builds. Or they plan
specifically not to have such dependencies to make automation easier.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are su
the jobs that run on particular nodes - and ideally, a stable
repository where the results can be stored and interchanged so you
don't have to count on the node that did the job still being available
when you want the results from it.
--
Les Mikesell
lesmikes...@gmail.com
--
You received thi
same FS root? I wouldn't expect
that to work at all, at least not concurrently.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and stop receiving emails
nkins do it that way from the start (at least
allowing the job definition to be stored/versioned as part of the
project)?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from
he brain melt if I start explaining them now ;-)
This thread probably isn't the right place, but I've always preferred
the 'think first, then act' approach and would like to see that
explanation.In particular, how direct/complete is the relationship
between what you can pass
fined
> in the literate build... I may find ways to pull them in but it needs
> thinking or the solution will uglify
Abstractions are nice as long as the person who wrote them anticipated
exactly what you need to describe. But somehow that isn't always the
case, so having a way to work
was working pretty well under 1.456, so
> what has changed since then ?
How many different versions of the slave.jar were you running then?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" gro
in other projects). Can I describe the layout of that
tree such that it might be different in a branch build than in other
versions?That may not be a reasonable thing to want to do, but
that's not quite the point here.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this
s some conflict (and i
> suspect the default port) when i launch two slaves, even with two different
> masters
I thought you said it only happens when you start with your custom ssh
script, not the internal ssh support. The ports used shouldn't
change.
--
Les Mikesell
lesmikes
especially when they need
different jars of the same name?Perhaps more to the point, why do
you want to do this? Maybe there is a better approach to whatever it
is you are trying to do.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed
ocs:
http://javadoc.jenkins-ci.org/
Or if you like brute force, you can download the source and feed it to
something like OpenGrok. And speaking of OpenGrok, has anyone set up
jenkins jobs to keep OpenGrok views of source up to date?
--
Les Mikesell
lesmikes...@gmail.com
--
You receiv
s for your VCS
checkout. At least with subversion you get some choices about how
much to keep/update between runs. If you have multiple slaves,
Jenkins will normally use the one where the job was built previously
but various circumstances can make it pick a different on.
--
Les Mikesell
possible.
OpenGrok is a little more computer brute force and a lot more
human-friendly. It does a cross reference and indexes that and the
full text with lucene to give you a color-coded source browser (for
multiple languages) with near-instant searches and automatic links to
jump between defi
nished?
> I can see that Junit can display report contents.
>
Can't you do this with a post-build step of 'archive the artifacts'?
Or do you want them to show up somewhere else?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subsc
ld happens on a different slave.
--
Les Mikesell
lesmikes...@gmail.com
>
> --
> 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
nterpreter line.
> If I run the same script through a terminal, it will work.
But probably not if you type:
sh scriptname
> Please help me to resolve this.
The simple way is to have an actual shell script that executes your
expect script. But using ssh keys instead of a password
nd usually should) apply a
label to a set of nodes and restrict builds to nodes with that label,
but the only preference ordering within the group is that jenkins will
use the node that did the last build if it is available - otherwise, I
think it picks at random from the others with the sp
here an ant located in the PATH on the machine?
--
Les Mikesell
lesmikes...@gmail.com
--
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 em
ave and use' not 'pick
something random' in this case. Can't you just specify some version
number that you know is available and will work?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenk
know ahead of time
instead of building at runtime?If it has to do with component and
library versions, we handle that with svn externals too - controlling
the versioning by referencing tags.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscrib
ax available in a normal jenkins job that could poll an scm and
collect final results. Haven't looked at scripler, though. Does it
provide all the same operations and compact notation?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribe
step. I haven't looked at scriptler yet - if it has a simple way to
run other jobs in parallel and access their results that may be what I
need.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins U
so, for example, the bash shell would source
~/.bashrc, but not /etc/profile, ~/.bash_profile or ~./.profile.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group
If could have a way to have the initial repository be kept uptodate with
> master repo that could work too - but didn't find reliable way to do that.
>
> Any hints ?
How about a scheduled (cron or jenkins job) 'get fetch' in the repo
that you cloned from the master?
t me that would tell me the best path
> for growing my Jenkins "farm"?
Just add slaves. Give the slaves labels so you can control where
things run by groups rather than individual nodes, and set 'restrict
where job can run' to an appropriate label in each job.
--
Les Mi
(and any other apps that need mail support) to
deliver to localhost with smtp on port 25. As a side effect you'll
get queuing and retries if the hub is not available.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Grou
ick-fix for programs that can't disable sleeping themselves.
--
Les Mikesell
lesmikes...@gmail.com
--
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, se
directly) guarantees that
> /usr/local/well-known-location/env.properties is a complete file at any time.
Doesn't that invite a race condition if multiple jobs run
concurrently? (Assuming they need different values...),
--
Les Mikesell
lesmikes...@gmail.com
--
You received
roovy plugin to
figure out where each instance is running and change the options
appropriately, but for a small number of targets it is probably easier
to just create new jobs as copies of the initial version and edit the
target label and build command.
--
Les Mikesell
lesmikes...@gmail.
And how would you
get the production and debug builds for each arch to run concurrently
on different nodes - where again it is just compiler options that
differ?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
&quo
d everywhere. If you use some conventions for which node
performs which type of build you might make the same batch file work
without too much trouble.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins U
he revision numbers.
Not sure even this will work for the svn externals unless they are
pointing to tags or peg revisions.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe fro
x27;d recommend separately tagging
the components you are pulling in separately. That is, use tags or
peg revisions in your svn externals and whatever you use to check out
from git. That way each component can have its own
development/release scheduling without breaking your main programs
that use the
ly the same command on each
node, so look at the xshell and groovy plugins for ways to make that
command do the right things on different operating systems.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"J
this specific case (or even how they handle java in
general), but it is common for Red Hat to backport fixes and updates
without changing the base version numbers on their packages so it
fairly hard to match up bugs with versions in RHEL/Centos unless you
look through the RPM changelog.
-
the master, do they run in the main jvm
instance? I always run builds on slaves which might help with never
running out of memory.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To
it through
> the interface.
>
> any ideas ?
I think the script console can only connect to one node at a time.
You could put you command in a matrix (multi-configuration) job and
select all the nodes, though.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because
ipt console? If the target is windows you may have to keep
in mind that some of the things you think are commands are build into
the cmd program.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins U
I, though.
You could, of course run a script on the master that uses ssh (with
keys configured for passwordless access) to run the commands remotely
on the slaves without very much jenkins involvement. That would be a
more natural approach for me since I use ssh much more than java or
groovy.
--
eady been built. Or, if you want more
complex control of multiple jobs you might use the build flow plugin.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group a
oving them from the jenkins configuration and having to
add them back when needed.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and stop receiving emails f
:XX:XX] - Request URL / not found - doesn't
>> match any webapp prefix
>
> So I am not sure whether my apache2 is configured wrong or whether this is a
> problem with jenkins? I really appriciate any help!
Adding --prefix=/jenkins to the jenkins config moves the url from the
default
protect your files from concurrent access. Normally you'd want
independent checkouts of the source and 'archive artifcacts' to
consolidate the results back on the master.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to
ng for
read-only access to a common set of tools and libraries, but write
everything locally - on VMs the workspace is on a separate volume that
can be rebuild as needed instead of having to expand a disk in place.
Even if we used network mapped workspace, I'd keep it unique per
slave to
he same with manual checkouts and builds.
--
Les Mikesell
lesmikes...@gmail.com
--
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-u
on
control system so it will automatically be checked out on the build
slave? And you can update it the same way as the source.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
T
und which also seems cumbersome.I think it
would be nice to have the high-level methods for common operations -
at least the things build flow would do - in a groovy library that
would be usable directly without having to know all the details and
that would work as an ordinary build step inside a job.
nd out, that if i set all slave by "deploy on demand", jenkins is alway
> just deploying on the master.
>
> What could be the reason?
Are you using 'restrict where this job can run' in the job? With
labels or node names?
--
Les Mikesell
lesmikes...@gmail.
w a custom workspace would relate to the choice
of nodes. That should be relative to the root filesystem of the
already-selected slave. Did you also set the 'restrict where job can
run' option?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because yo
cuit, forwarding other queries to your upstream
resolvers.
--
Les Mikesell
lesmikes...@gmail.com
--
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,
t should work with node names too. You need something there if you
want to control where it runs.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and st
? Why not include the ant script in the checkout
and let jenkins do it in the right place for you?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and s
also performed by Ant on
>> > slave.
>> >
>> > I need the "poll SCM" to kick off the build process, for nothing else,
>> > i.e
>> > checkout of repository in workspace has no sense on my master.
>>
>> Why do it that way? Why not include
add a post-build action to 'archive the artifacts' and
specify the files you want.
--
Les Mikesell
lesmikes...@gmail.com
On Fri, Jan 17, 2014 at 9:35 AM, kgiloo wrote:
> Thank you so much for your explanation!
> It seems now easy to define a new slave, trigger svn
recently created release candidate.
>
> http://jenkins-ci.org/stable-rc
>
Is there a changelog for that?
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from t
I don't see similar problems with an up
to date CentOS 5.10 and the stock openjdk.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and stop receiving
t; configuration page.
>
Or, if you want to install it yourself you can give the version a name
in the global config, then in the per-node config, in node properties,
tool locations, add a path to find it on that node.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this messag
> mode afterwards :))
>
I sort-of manage it by installing versions on a shared drive that all
the slaves see - windows slaves use UNC paths and linux slaves have an
nfs mount in a common location.So there is some extra work, but
not a full install on every target.
--
Les Mikes
staying up to date on general principles, especially
with redhat or centos where the updates rarely break things.But,
for jenkins the jvm version will likely make more difference than the
OS, and even that isn't too critical.
--
Les Mikesell
lesmikes...@gmail.com
--
You rec
environment?
Generally you shouldn't rely on local environment settings on slaves
since these may be different from a logged-in session. If you need
something to be set you can do it in the jenkins per-node config.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message
ild triggers' check 'build periodically' and fill in the
schedule with the cron-like syntax that is explained if you click the
? symbol. Something like 0 1,13 * * * would run at 1 am and 1 pm.
Then add a build step to execute a windows batch command to do the work.
Did you try th
in workspace
> C:\Users\lily.fu\.jenkins\jobs\LilyTest_LiquibaseDiff\workspace
> [workspace] $ sh -xe C:\windows\TEMP\hudson2336063048912468753.sh
That looks like you picked a build step of 'execute shell'
(appropriate for unix/linux) instead of 'execute windows batch
command' wh
On Tue, Feb 11, 2014 at 10:23 AM, David Aldrich
wrote:
>
> Please can anyone suggest why a job, running on a Windows slave, might fail
> part way through with error "No route to host"?
>
Network problem?
--
Les Mikesell
lesmikes...@gmail.com
--
You received th
1 - 100 of 468 matches
Mail list logo
