Another option if you want to do it programmatically could be to use an API
call via url hit to take machines offline, then put them back online. The
general format of the call should be available pretty easily with a web
search, but we use something like this in a curl call when wanting to take
m
Sean,
Without knowing your specifics of memory capacity, number of current
executors, executor usage, etc., it's hard to give any hard guidelines.
Perhaps try to reduce by 25% the number of executors and see what happens.
There's no magic number as each installation is different.
Scott
On Tue,
If I were to set an environment variable at both the server and node level,
is there a guaranteed order of precedence for which would be actually used
for a build? I have a global variable which I wish to override for only a
single node. My brief testing indicates that the node variable is what
g
We have a build matrix job which runs on a regular basis, and since we have
it configured to be able to run concurrent builds at the same time, often
times there is more than one build running at a time. This build is set to
use Git for source control, so the parent build first pulls code down, th
Lars,
Check to make sure you have the same plugins installed/active on the target
system as you do on the source configuration (also make sure the versions
of the plugins are the same or at least as close as possible). Also, check
to make sure file owner/permissions are correct on the target syst
A weird thing has started happening for us, where our Jenkins instance (run
out of Tomcat) is coming up both at http: (port 80) and https: (port 443).
In our tomcat server.xml we do have a redirect as follows:
And this hasn't changed for months, though the behavior of also responding
on port
I believe that Jenkins is is set to not put another build in the queue if
one is already there, waiting to be run, if it is triggered manually by an
upstream build. I do not know if this can be overridden in some way or
not. I know in our instances, a Gerrit patch set that triggers a build can
pu
Check to make sure permissions/owner settings on the Ubuntu side are
correct. To see what the correct ones should be, the easiest way would be
to create a new dummy job on the Ubuntu side to see what gets created.
Also check the Jenkins log to make sure you are not missing any plugins and
that mi
Alex,
IF your builds are doing a lot of disk reading/writing and you're running
multiple builds in parallel, you might be limited in build performance by
how fast your hardware can read/write to disk. If multiple builds are all
doing a bunch of disk activity at the same time, you may be maxing ou
Alex,
I'm not a VMWare admin, so your experience would probably be better than
anything I could suggest for setting up VMware nodes. Is your build set up
(such as with a multiconfiguration / matrix build) to use all 3 nodes for
its work? If your build already uses all 3 slaves and all 3 are maxi
Alex,
The first question I'd ask is what part of the build is running at 100%
cpu, and can you tell what processes are consuming the cpu at that time?
You need to figure out the bottlenecks and where your pain points are
first, before trying to solve your performance issues (or perceived
issues).
We have an established Jenkins instance where I'm trying to set up a new
node, which I've done many times before. However, previously when I've set
up a new node it's given the headless command line to use of the format:
java -jar slave.jar -jnlpUrl
https://hostname:port/computer/nodename/slave-a
Ed,
I would guess you have your Jenkins server configured to be considered at
127.0.0.1 (i.e. loopback) and so any referece to it from other machine
(i.e. slaves) are trying to hit 127.0.0.1 for the URL which (since it's the
loopback address) is trying to contact to itself rather than the master.
I am working on a project where I want to try and programmatically obtain a
specific node's Jenkins environment variable via an api call (i.e.
/computer//api/xml) that's set as a node property
environment variable on the node configure page. However, I cannot find
any way to get the /api/xml to re
Gregg,
Sounds like a very bad idea to me. As I have experienced, Jenkins caches
project information in memory and blindly writes out configuration
information to disk for restart purposes, and typically only reads disk
data in case of a Jenkins restart. So, you could very easily end up in a
situ
We've seen this occasionally on Ubuntu as well with recent versions.
Scott
On Tue, Aug 12, 2014 at 3:41 PM, reignmang wrote:
> I'm having the same issue with LTS 1.565.1 on Win7 64 bit. Did you figure
> out a solution??
>
> On Wednesday, July 30, 2014 4:11:29 PM UTC-5, protomouse wrote:
>>
>>
nfirm
that emails always get routed back to the master to send.
Scott
On Mon, Jul 28, 2014 at 11:05 AM, Slide wrote:
> I believe it gets executed on the individual build job nodes.
>
>
> On Mon, Jul 28, 2014 at 8:14 AM, Scott Evans wrote:
>
>> Does anyone know offhand where
Does anyone know offhand where the Mailer plugin gets executed for a
build? Does it get executed on the master Jenkins node, the
multi-configuration parent, or the individual child build job nodes? We're
trying to limit as much as possible which nodes in our system need to have
access to our smtp
Rob, have you tried running in different browsers? One in Chrome and the
other in Firefox? I've not had luck running such a setup in a single
browser as the authentication seems to carry from tab to tab, but I've had
success running different browsers to achieve this result.
Scott
On Thu, Jul
Kevin,
I don't do my windows slaves by service, I do the java start on them, so I
don't know what you'd see/experience if you tried to start more than one as
a service.
Scott
On Wed, Jul 16, 2014 at 2:18 PM, Kevin wrote:
> Also, when start 2nd slave, do we suppose to see two "Jenkins Slave" i
What you'd need to do is set up multiple agents on the slave node, one
agent for each Jenkins master. You would also need to set up independent
workspaces for each agent as well. This is assuming that your test
hardware can handle more than one job at a time. We do this a lot and it
works fine,
Gavin, I ended up writing a custom monitoring routine which pulls the api
information from Jenkins and sends me a notification if a job is running
long. There's a flag in the api xml that tells my script if a job is
potentially stuck that I key off of.
I run it on a separate box and just use curl
Bill,
We ended up restricting our project naming convention to not allow spaces
in the project names, just for reasons like this - tools which don't
understand spaces in path names were causing significant headaches and
script coding workarounds that weren't worth the effort being expended.
Scott
Hi!
I've successfully implemented the use of a url call to toggle the status of
a build note to either be online, or offline:
/computer//toggleOffline?offlineMessage=Offline
Which works fine. However, each call of this URL just toggles the current
status. Is there another parameter that I coul
delabel parameter plugin but i am getting IO
> exception, unable to install this plugin
> Is this proper plugin which I installed? or any other plug in is there...
> Please can you provide sample tutorial where I can see the proper info.
>
> Thanks in advance Evans.
>
> Regards
Panikera,
Look into how to use labels, as that's exactly their purpose.
Scott
On Tue, May 13, 2014 at 8:19 AM, Panikera Raj wrote:
> Hi All,
>
> I am using master and Slave machanism, Assume that Master in busy with
> build generation, now if I got new request comes I have to distribute the
>
Amol,
This seems rather unusual to need to have a job run every 5 seconds. If
you could let us know what you're trying to accomplish, there is most
likely another way that's better than a continuously running job.
Scott
On Wed, May 7, 2014 at 10:03 AM, Daniel Beck wrote:
> This isn't straigh
Make sure the service is set to interact with the desktop, but I believe
you also need to make sure that there's an account logged in, so there's a
desktop session to interact with.
Scott
On Thu, May 1, 2014 at 9:57 AM, wrote:
> I'm not sure about that. Is there a reason you need to use a cus
We have a bunch of matrix (multi-configuration) build jobs which are
configured to use a Git plugin to pull source code into the job workspace.
Unfortunately, what happens is that the parent job pulls the full set of
code and expands it into the parent workspace, only to then fire off all of
the ch
What is the load on the server while this is going on? Is the Jenkins java
process doing anything that's using up significant CPU?
Scott
On Apr 30, 2014 12:57 PM, "Seth Floyd" wrote:
> Nope...that didnt help. Its weird how its just on a few pages that this
> happens. I can get to the configurat
My guess is that you're running the slave agent as a service and it doesn't
have access to network resources. When you run it manually you're logged
into the box and your account probably has network access privileges.
Scott
On Tue, Apr 29, 2014 at 3:14 PM, wrote:
> On my slave machine, which
Matthew,
You're correct that it's an unlikely scenario with dhcp, but it was the
first thing I thought of. The next time it fails, not only look at the
agent log, but look into the server logs and see if there's anything there
that might shed light on what's going on.
Is it always the same VM th
Matthew, just a quick thought that's not 100% related to your issue, but
is the first thing that came to mind when you said you're having issues
with a slave upon startup. Are these dhcp machines by chance, and is it
possible that this slave is somehow getting an ip address when it wakes up
of a
Kolos,
A few quick thoughts...
Is this a windows slave, and does this build have some aspect of it that
needs to interact with the console (i.e. pop-up window, etc.) that it
doesn't have permissions to do so (or is it waiting for some user input?)?
Does it always hang at the same spot? Based on
d this confuses me. I dont understand why? I"m clearly typing
> the correct path. Maybe I should create the storage path local to the
> workspace?
>
>
> On Tue, Apr 22, 2014 at 9:25 AM, Scott Evans wrote:
>
>> Chanda,
>>
>> I don't use the CC plugin, but
Chanda,
I don't use the CC plugin, but it looks like the call it's doing to try and
create your local view is failing on the view storage directory path
format. In your log, it looks like there's "null" that's getting put in,
where a UNC-style path likely should be. I assume you're already fami
To ???:
A few quick thoughts (since I don't have access to your machine and
environment):
- You said you ran the command-line job on the Jenkins server itself? Is
this the same physical box where the build ran that's causing problems?
Most people are set up with a Jenkins server and then agent m
Don't know if it's relevant to what's going on in your situation (as we
don't have our Win7 agents running as services, but do you have your Win7
service for Jenkins set to be able to interact with the desktop? (I don't
remember specifically where that's set or what it's called, but I know we
had
I've found in many cases that I've needed to bounce the Jenkins process
(both the master and slaves) to pick up changed environment variables.
Can't guarantee that's what's going on here, but is worth trying if you
haven't already done so.
Scott
On Mon, Mar 10, 2014 at 2:40 PM, Christian Willman
We have a developer who would like to kick off a build if either of two
conditions is met (such as a Git change for one condition and a file change
as a second dependency. He has things set up to run the build
conditionally if either condition is met, but what he's finding is that if
both conditio
AndInvoke(Function.java:151)
at
org.kohsuke.stapler.Function.bindAndInvokeAndServeResponse(Function.java:90)
at org.kohsuke.stapler.MetaClass$1.doDispatch(MetaClass.java:111)
at
org.kohsuke.stapler.NameBasedDispatcher.dispatch(NameBasedDispatcher.java:53)
at org.kohsuke.stapler.Stapler.tryInvoke(Stapler.java:677)
... 73 more
On Tue, Jan 2
Daniel - Thanks! Exactly what I was needing (and missing in reading the
documentation).
Scott
On Tue, Jan 28, 2014 at 3:49 PM, Daniel Beck wrote:
>
> On 28.01.2014, at 22:45, Scott Evans wrote:
>
> > 1. Is this expected behavior?
>
> Yes.
>
> > 3. If this is
I'm trying to roll my own external monitoring for my Jenkins server using
PHP to access the Jenkins API and then parse the returned XML. I'm able to
(using curl) pull out XML from Jenkins using the following URL format:
https:///computer//api/xml
But although I am getting back a bunch of XML dat
Eric,
We use that all the time. Under Build Environment for each job there's a
click box for "Inject password to the build as environment variables" and a
sub-check for Global Passwords. It should be masked in any log output, but
available to each job that's configured to use global passwords.
Is there any way to have a column on the main screen which shows (and
allows sorting) based on the time since the last build? There are already
columns available for the last success and last failure, but no columns
seem to be available to show the last build, no matter the result? We have
some b
Eric,
If you don't care the order that they run, just that they don't run
simulatnously, my first thought is to set up a build executor with only one
agent, and assign this job to that single executor. That way, only one can
run at a time. There's also a build throttling plug-in, but I haven't
l
Is there a way to set a jenkins job to fail a multiconfiguration job if one
of its children have been sitting in the queue for more than x amount of
time?
I have set up a multiconfiguration job to run maintenance on every node in
our system on a regular basis. However, if a node is down, its chil
Have you looked into using a matrix build to run a build job on all of the
slave nodes at the same time? Just put the script in a common location and
reference it via unc path in your build steps. Then set the matrix job to
run on each of the nodes and then run the job, should be pretty easy,
esp
I'm not familiar with any offhand which are illegal, though I'd suggest as
best practices to avoid anything that's not a standard alphabetic, numeric,
or dash/underscore. Depending on what platform(s) you're working on, I'd
suggest for sure to avoid spaces, as that just gets weird with some
comman
We use a post-build action to "Archive the artifacts" with an appropriate
file mask. I believe this is a standard Jenkins function and not via
plug-in. This will attach whatever items you want to the Jenkins build.
Scott
On Wed, Oct 9, 2013 at 3:06 PM, wrote:
> I cannot figure out if any of
We have a Jenkins server (1.518) running on Ubuntu, with many plugins
active, including "Jenkins GIT client plugin" 1.3.0 and "Jenkins GIT
plugin" 1.5.0. What we're seeing is that in many cases on Windows agents
of either Win7x64 or WinXP (not on Ubuntu), that often (but not always), a
Git .pack f
Has anyone else experienced a situation where on an occasional basis, a
build will fail with an "Ant not found" error? It's very sporadic, but
will occur very rarely, where a given build will run 25 times successfully,
then the next time on the same machine and same executor will die with this
err
If you're running Windows, do you have any sort of virus protection running
on your jenkins server box that might be blocking outgoing emails, or
possibly port 25 is being blocked on an outgoing request? I assume when
you say you can telnet to the email server machine that you're telnetting
to por
Have more than one agent running on a slave machine, one for each master,
and just make sure they're pointing to different work areas. You might end
up with two builds at a time running on a given box if both masters give it
a build at the same time.
Scott
On Thu, May 30, 2013 at 8:54 PM, Kamal
Dan,
I'm not a mac person so can't help you in detail, but can you, from your
mac, browse to your Jenkins server with a web browser? It almost sounds
like your slave mac isn't giving the jenkins slave process enough
permissions to reach out, or there is some sort of firewall blocking
connections
Just put a label on the agents and assign both the build and test jobs to
that label and it should work just as you expect. Note that unless you use
a matrix configuration, a job will only use one agent at a time, so you'd
want to look into matrix jobs if you have several sets of tests you want to
Currently, we have Jenkins set up to trigger on a Gerrit patchset upload or
merge to do a build, which works fine. We also use a build version plugin
to version the build with a custom build number sequence
(w.x.y.buildnumber). However, we now have been asked to implement having a
single Jenkins
Antonio,
Is this on a Windows box with more than one agent configured to run on it
and possibly more than one agent is trying to do a git pull at the same
time? If so, it may be ssh hanging on you. I recently went through this
myself, and found the following reference a great help.
http://darre
I see the same behavior and certainly assume it is by design. The
workspaces are likely still there and can be accessed directly if
you're on that node itself, just not through Jenkins.
Scott
On Tue, Mar 26, 2013 at 12:30 PM, Robert Moore wrote:
> This behavior seems new to me but perhaps I've
You might need to put intelligence into your build script to know which
platform it's on in order to then call the proper installer.
Scott
On Fri, Feb 1, 2013 at 3:39 PM, zw wrote:
> Hi All
>
> We like to use the multi-configuration project where the values on 1 axis
> has values windows and li
can put a regex to keep spaces out
>
> There is also the create job advanced plugin, which has a checkbox for
> turning spaces into dashes:
> https://wiki.jenkins-ci.org/display/JENKINS/Create+Job+Advanced+Plugin
>
> Will
>
> On 01/23/2013 03:40 PM, Scott Evans wrote:
>
&g
Does anyone have a suggestion on a way to programatically keep developers
from creating new projects which have spaces in the names of the projects?
Somehow, my proposal of just telling them "don't do it" and when they do it
anyway going an rapping their knuckles with a yardstick doesn't seem to fl
Slide,
Yes, that would do pretty much what I needed, if I weren't needing to run
this on a Win-7 x64 box which doesn't have command.com on it. In any
event, since my original email, we've decided that since it's a "bad thing"
to have spaces in project names (especially when we're multi-os), that
Within my ant build script, I have a legacy program that is failing because
the workspace name has a space in it, and so it's failing as the whole path
isn't getting passed in as an argument to this program. It really isn't a
solution at this time to change the project name to one without spaces,
Check to make sure your java memory settings didn't get set too low during
the upgrade, as we had issues with slow reponses at times which cleared up
once we configured the Jenkins java process to have more memory available
than the default. I believe this setting can be configured in jenkins.xml.
In your jenkins build, can you try just running a command "nslookup" (or
ping would probably work too) on your agent machine to see what response
you get from actually within the build session? That might indicate
whether you're even getting a valid lookup back from DNS from within your
job enviro
Not sure what OS you're running on, but could it be that the user that the
Jenkins agent is running as might not have permissions to access network
resources to do a DNS lookup? Seems odd, but that's the only thing that
comes to mind for this case.
Scott
On Thu, Dec 13, 2012 at 7:29 PM, mwpowell
The error would indicate to me that it can't find the host where your repo
is, whether a typo in that variable setting that's not getting resolved
through your nameserver, or some sort of network issue, most likely.
On Thu, Dec 13, 2012 at 3:47 PM, mwpowellhtx wrote:
> To clarify, I've edited
Looks like you're missing an environment setting for repo-host
Scott
On Thu, Dec 13, 2012 at 3:35 PM, mwpowellhtx wrote:
> stderr: ssh: : no address associated with name
The amount of memory you have physically installed is completely different
than the amount of memory the java process is allowed to allocate & use.
Changing the settings as outlined below will give your java process more
memory to work with.
Scott
On Thu, Dec 6, 2012 at 8:34 AM, El alaoui Mohamed
Steve,
Are you sure your ftp session has write permissions on the target
directory? Can you put files elsewhere on the target server (i.e. initial
directory when you first initiate the ftp session)? Can you list other
files in the target directory (either with an ls or dir command)?
One other q
What do you have for your java memory settings? Primarily -Xmx would
define the max amount of memory your main java process can allocate.
Scott
On Thu, Nov 15, 2012 at 3:10 PM, MDKF wrote:
> Can anyone help with this issue?
>
> On Monday, November 12, 2012 10:36:13 AM UTC-5, MDKF wrote:
>>
>>
We're slowly migrating to a Jenkins-based build system from an older system
based significantly on Sharepoint for data management. As part of that
system, I've written a significant number of external monitoring tools
which track build run times, build machine status, etc. by pulling data
from the
On the main builds dashboard, there are several columns, two of which are
Last Success and Last Failure. However, there is not a column which sorts
builds by last build, whether success or failure. In other words, I want
to be able to sort on which builds were most recently run, regardless of
whe
Ronan,
My guess is that you have the agent configured to run as a service, rather
than a java process on the desktop. If so, you'll need to set that service
to run as a specific user rather than local system account. That can be
found under the LogOn tab of the jenkins service in question.
Scot
Ed,
Based on my experience, it will purge old builds (by date or quantity) only
when a new build of that type completes. In your case, once you run one,
it'll should automatically delete the 91 "extra" builds which are no longer
within the retention policies for that build type. Note that it wil
I assume you're seeing these errors on your agent box. What are your
java memory settings on that box? I'm guessing it's a problem on the
agent side and not the server.
Scott
On Tue, Aug 28, 2012 at 12:26 PM, Vishal Bansod wrote:
> Hi all,
>
>
> Recently I started experiencing a problem below
e is some resource contention taking place.
>
> Again from my log snippet I really feel there is an issue spawning the
> powershell session.
>
> -barrett
>
> On Thu, Aug 9, 2012 at 11:10 PM, Scott Evans wrote:
>>
>> Just a thought - are the successes and failure
Just a thought - are the successes and failures both running on the
same agent box? If not, then perhaps one is configured differently
than the other.
Scott
On Thu, Aug 9, 2012 at 9:04 PM, j.barrett strausser
wrote:
> That's absolutely the reason it is being marked as failure. Jenkins
> interpr
79 matches
Mail list logo