Multiple branches best practice ?

classic Classic list List threaded Threaded
28 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Re: SCM change when there is no change

Kedar Mhaswade-3
On 09/27/2010 02:52 PM, Richard Bywater wrote:
> Just to throw another SCM type into the mix, I'm also seeing this on
> Clearcase - unfortunately because the polling logs aren't kept (I
> don't think?) then I can't tell why it kicked off multiple builds.
>
> Perhaps the polling log mechanism could be changed to keep the logs
> for a period to help solve issues like this?
>    
I think the polling logs are (supposed to be) kept in the file
<Hudson-Home>/jobs/<job-name>/scm-polling.log when SCM trigger
is configured. (I find that at least in the case of subversion).

Are you saying you don't find that file on disk?

Also, interestingly, the log output that Thomas sent says:
"no change for ...", so it appears that Hudson knows there is no change
in SCM
and still, a new build is scheduled and run.

> Richard.
>
> On 9/28/10, Curt Patrick<[hidden email]>  wrote:
>    
>> I see the same thing, rather frequently, but against a Perforce repository.
>> So I doubt seriously that this can be blamed on anything Subversion
>> specific.
>>
>> Curt
>>
>> -----Original Message-----
>> From: Thomas Ferris Nicolaisen [mailto:[hidden email]]
>> Sent: Monday, September 27, 2010 6:34 AM
>> To: [hidden email]
>> Subject: Re: RE: SCM change when there is no change
>>
>> We're having the same problem here occasionally (like, once a month or
>> so, 20 developers collaborating on various projects). Suddenly Hudson
>> will trigger a bunch of builds in different projects with this output
>>
>> Started by an SCM change
>> Building remotely on [...]
>> Updating https://scm.company.com/svn/java/service/trunk
>> At revision 102945
>> no change for https://scm.company.com/svn/java/service/trunk since the
>> previous build
>> Parsing POMs
>> ....
>>
>> One theory is that it something to do with instability in the
>> Subversion server's availability. I noticed that our mail server was
>> undergoing some maintenance at the same time, and maybe a techie
>> pulled the plug on the Subversion server at the same time. I haven't
>> observed exact correspondence, it's just a shot in the dark.
>>
>> Running Hudson 1.368, the servers are running stable Debian.
>>
>>
>> On Wed, Aug 4, 2010 at 1:15 AM, Sri Bolle<[hidden email]>  wrote:
>>      
>>> I am not sure what versions you are using.. I had similar issues with some
>>> of the hudson core and perforce plugin during testing... now using 1.366
>>> and
>>> latest perforce plugin with no issues...
>>>
>>> On Fri, Jul 30, 2010 at 12:52 PM, Curt Patrick<[hidden email]>
>>> wrote:
>>>        
>>>> Here is another bit of oddity about the same job: �In the console output
>>>> it says that the build finished successfully after about 2 minutes (which
>>>> is
>>>> what I'd suspect), but �hours later the job is "still running".
>>>> �Something
>>>> is really stuck.
>>>>
>>>> I doubt that this is related to the original problem I reported since
>>>> I've
>>>> seen the "no change" often in cases where the build went fine, but I
>>>> thought
>>>> I'd mention it.
>>>>
>>>> Thanks for your attention.
>>>>
>>>> Curt
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: Neerenberg, Aaron [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:46 AM
>>>> To: [hidden email]
>>>> Subject: RE: RE: SCM change when there is no change
>>>>
>>>> Not really. �I was wondering if the polling output might have just not
>>>> saved into the change set correctly, but that doesn't seem to be the
>>>> case.
>>>>
>>>> Regards,
>>>> Aaron
>>>>
>>>> -----Original Message-----
>>>> From: Curt Patrick [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:39 AM
>>>> To: '[hidden email]'
>>>> Subject: RE: SCM change when there is no change
>>>>
>>>> Here is the entire log in all its glory. �Does it give you some clue
>>>> that I'm missing?:
>>>>
>>>> Started on Jul 30, 2010 9:37:39 AM
>>>> Looking for changes...
>>>> Using remote perforce client:
>>>> hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>> [OE-HUBS-Oasis-Java-dev-build] $ p4 workspace -o
>>>> hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>> Saving modified client hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>> [OE-HUBS-Oasis-Java-dev-build] $ p4 -s client -i Last sync'd change was
>>>> 398604 [OE-HUBS-Oasis-Java-dev-build] $ p4 changes -m 2
>>>> //hudson-OE-HUBS-Oasis-Java-dev-build--1273170058/...
>>>> Latest submitted change selected by workspace is 398604 Assuming that
>>>> the workspace definition has not changed.
>>>> Done. Took 0.55 sec
>>>> No changes
>>>>
>>>> -----Original Message-----
>>>> From: Neerenberg, Aaron [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:35 AM
>>>> To: [hidden email]
>>>> Subject: RE: SCM change when there is no change
>>>>
>>>> Does anything show in the job's Polling log?
>>>>
>>>> -----Original Message-----
>>>> From: Curt Patrick [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:33 AM
>>>> To: '[hidden email]'
>>>> Subject: SCM change when there is no change
>>>>
>>>> Hello,
>>>>
>>>> Can anyone explain what it means when a job says it is "Started by an
>>>> SCM change" but there is nothing reported in the "Summary of Changes"
>>>> section? �This happens fairly often and puzzles me. �BTW, I'm using
>>>> Perforce.
>>>>
>>>> Curt
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>          
>>>
>>>        
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>>      
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>    


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

RE: SCM change when there is no change

Curt Patrick
When you look at the page for the build in P4V it says "No changes from the last build".  And below that it says "Started by an SCM change".  That sure doesn't make much sense.  I asked what that meant some time ago and got no useful feedback.

Richard may be saying something different than I am, but my problem with the polling log is that, every time, by the time I realize there is a problem and look at the log it has already polled again.  So I can never catch a polling problem "in the act".

Curt

-----Original Message-----
From: Kedar Mhaswade [mailto:[hidden email]]
Sent: Monday, September 27, 2010 3:38 PM
To: [hidden email]
Subject: Re: SCM change when there is no change

On 09/27/2010 02:52 PM, Richard Bywater wrote:
> Just to throw another SCM type into the mix, I'm also seeing this on
> Clearcase - unfortunately because the polling logs aren't kept (I
> don't think?) then I can't tell why it kicked off multiple builds.
>
> Perhaps the polling log mechanism could be changed to keep the logs
> for a period to help solve issues like this?
>    
I think the polling logs are (supposed to be) kept in the file
<Hudson-Home>/jobs/<job-name>/scm-polling.log when SCM trigger
is configured. (I find that at least in the case of subversion).

Are you saying you don't find that file on disk?

Also, interestingly, the log output that Thomas sent says:
"no change for ...", so it appears that Hudson knows there is no change
in SCM
and still, a new build is scheduled and run.

> Richard.
>
> On 9/28/10, Curt Patrick<[hidden email]>  wrote:
>    
>> I see the same thing, rather frequently, but against a Perforce repository.
>> So I doubt seriously that this can be blamed on anything Subversion
>> specific.
>>
>> Curt
>>
>> -----Original Message-----
>> From: Thomas Ferris Nicolaisen [mailto:[hidden email]]
>> Sent: Monday, September 27, 2010 6:34 AM
>> To: [hidden email]
>> Subject: Re: RE: SCM change when there is no change
>>
>> We're having the same problem here occasionally (like, once a month or
>> so, 20 developers collaborating on various projects). Suddenly Hudson
>> will trigger a bunch of builds in different projects with this output
>>
>> Started by an SCM change
>> Building remotely on [...]
>> Updating https://scm.company.com/svn/java/service/trunk
>> At revision 102945
>> no change for https://scm.company.com/svn/java/service/trunk since the
>> previous build
>> Parsing POMs
>> ....
>>
>> One theory is that it something to do with instability in the
>> Subversion server's availability. I noticed that our mail server was
>> undergoing some maintenance at the same time, and maybe a techie
>> pulled the plug on the Subversion server at the same time. I haven't
>> observed exact correspondence, it's just a shot in the dark.
>>
>> Running Hudson 1.368, the servers are running stable Debian.
>>
>>
>> On Wed, Aug 4, 2010 at 1:15 AM, Sri Bolle<[hidden email]>  wrote:
>>      
>>> I am not sure what versions you are using.. I had similar issues with some
>>> of the hudson core and perforce plugin during testing... now using 1.366
>>> and
>>> latest perforce plugin with no issues...
>>>
>>> On Fri, Jul 30, 2010 at 12:52 PM, Curt Patrick<[hidden email]>
>>> wrote:
>>>        
>>>> Here is another bit of oddity about the same job: �In the console output
>>>> it says that the build finished successfully after about 2 minutes (which
>>>> is
>>>> what I'd suspect), but �hours later the job is "still running".
>>>> �Something
>>>> is really stuck.
>>>>
>>>> I doubt that this is related to the original problem I reported since
>>>> I've
>>>> seen the "no change" often in cases where the build went fine, but I
>>>> thought
>>>> I'd mention it.
>>>>
>>>> Thanks for your attention.
>>>>
>>>> Curt
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: Neerenberg, Aaron [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:46 AM
>>>> To: [hidden email]
>>>> Subject: RE: RE: SCM change when there is no change
>>>>
>>>> Not really. �I was wondering if the polling output might have just not
>>>> saved into the change set correctly, but that doesn't seem to be the
>>>> case.
>>>>
>>>> Regards,
>>>> Aaron
>>>>
>>>> -----Original Message-----
>>>> From: Curt Patrick [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:39 AM
>>>> To: '[hidden email]'
>>>> Subject: RE: SCM change when there is no change
>>>>
>>>> Here is the entire log in all its glory. �Does it give you some clue
>>>> that I'm missing?:
>>>>
>>>> Started on Jul 30, 2010 9:37:39 AM
>>>> Looking for changes...
>>>> Using remote perforce client:
>>>> hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>> [OE-HUBS-Oasis-Java-dev-build] $ p4 workspace -o
>>>> hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>> Saving modified client hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>> [OE-HUBS-Oasis-Java-dev-build] $ p4 -s client -i Last sync'd change was
>>>> 398604 [OE-HUBS-Oasis-Java-dev-build] $ p4 changes -m 2
>>>> //hudson-OE-HUBS-Oasis-Java-dev-build--1273170058/...
>>>> Latest submitted change selected by workspace is 398604 Assuming that
>>>> the workspace definition has not changed.
>>>> Done. Took 0.55 sec
>>>> No changes
>>>>
>>>> -----Original Message-----
>>>> From: Neerenberg, Aaron [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:35 AM
>>>> To: [hidden email]
>>>> Subject: RE: SCM change when there is no change
>>>>
>>>> Does anything show in the job's Polling log?
>>>>
>>>> -----Original Message-----
>>>> From: Curt Patrick [mailto:[hidden email]]
>>>> Sent: Friday, July 30, 2010 9:33 AM
>>>> To: '[hidden email]'
>>>> Subject: SCM change when there is no change
>>>>
>>>> Hello,
>>>>
>>>> Can anyone explain what it means when a job says it is "Started by an
>>>> SCM change" but there is nothing reported in the "Summary of Changes"
>>>> section? �This happens fairly often and puzzles me. �BTW, I'm using
>>>> Perforce.
>>>>
>>>> Curt
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>          
>>>
>>>        
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>>      
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>    


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]


Reply | Threaded
Open this post in threaded view
|

Re: SCM change when there is no change

Richard Bywater
I'm on the same page as Curt re the polling logs. Yes I can see the
"current" polling log but once the system has polled again, the log is
overwritten and therefore one can't go back and try and troubleshoot
the multiple builds.

Richard.

On 9/28/10, Curt Patrick <[hidden email]> wrote:

> When you look at the page for the build in P4V it says "No changes from the
> last build".  And below that it says "Started by an SCM change".  That sure
> doesn't make much sense.  I asked what that meant some time ago and got no
> useful feedback.
>
> Richard may be saying something different than I am, but my problem with the
> polling log is that, every time, by the time I realize there is a problem
> and look at the log it has already polled again.  So I can never catch a
> polling problem "in the act".
>
> Curt
>
> -----Original Message-----
> From: Kedar Mhaswade [mailto:[hidden email]]
> Sent: Monday, September 27, 2010 3:38 PM
> To: [hidden email]
> Subject: Re: SCM change when there is no change
>
> On 09/27/2010 02:52 PM, Richard Bywater wrote:
>> Just to throw another SCM type into the mix, I'm also seeing this on
>> Clearcase - unfortunately because the polling logs aren't kept (I
>> don't think?) then I can't tell why it kicked off multiple builds.
>>
>> Perhaps the polling log mechanism could be changed to keep the logs
>> for a period to help solve issues like this?
>>
> I think the polling logs are (supposed to be) kept in the file
> <Hudson-Home>/jobs/<job-name>/scm-polling.log when SCM trigger
> is configured. (I find that at least in the case of subversion).
>
> Are you saying you don't find that file on disk?
>
> Also, interestingly, the log output that Thomas sent says:
> "no change for ...", so it appears that Hudson knows there is no change
> in SCM
> and still, a new build is scheduled and run.
>
>> Richard.
>>
>> On 9/28/10, Curt Patrick<[hidden email]>  wrote:
>>
>>> I see the same thing, rather frequently, but against a Perforce
>>> repository.
>>> So I doubt seriously that this can be blamed on anything Subversion
>>> specific.
>>>
>>> Curt
>>>
>>> -----Original Message-----
>>> From: Thomas Ferris Nicolaisen [mailto:[hidden email]]
>>> Sent: Monday, September 27, 2010 6:34 AM
>>> To: [hidden email]
>>> Subject: Re: RE: SCM change when there is no change
>>>
>>> We're having the same problem here occasionally (like, once a month or
>>> so, 20 developers collaborating on various projects). Suddenly Hudson
>>> will trigger a bunch of builds in different projects with this output
>>>
>>> Started by an SCM change
>>> Building remotely on [...]
>>> Updating https://scm.company.com/svn/java/service/trunk
>>> At revision 102945
>>> no change for https://scm.company.com/svn/java/service/trunk since the
>>> previous build
>>> Parsing POMs
>>> ....
>>>
>>> One theory is that it something to do with instability in the
>>> Subversion server's availability. I noticed that our mail server was
>>> undergoing some maintenance at the same time, and maybe a techie
>>> pulled the plug on the Subversion server at the same time. I haven't
>>> observed exact correspondence, it's just a shot in the dark.
>>>
>>> Running Hudson 1.368, the servers are running stable Debian.
>>>
>>>
>>> On Wed, Aug 4, 2010 at 1:15 AM, Sri Bolle<[hidden email]>
>>> wrote:
>>>
>>>> I am not sure what versions you are using.. I had similar issues with
>>>> some
>>>> of the hudson core and perforce plugin during testing... now using
>>>> 1.366
>>>> and
>>>> latest perforce plugin with no issues...
>>>>
>>>> On Fri, Jul 30, 2010 at 12:52 PM, Curt Patrick<[hidden email]>
>>>> wrote:
>>>>
>>>>> Here is another bit of oddity about the same job: �In the console
>>>>> output
>>>>> it says that the build finished successfully after about 2 minutes
>>>>> (which
>>>>> is
>>>>> what I'd suspect), but �hours later the job is "still running".
>>>>> �Something
>>>>> is really stuck.
>>>>>
>>>>> I doubt that this is related to the original problem I reported since
>>>>> I've
>>>>> seen the "no change" often in cases where the build went fine, but I
>>>>> thought
>>>>> I'd mention it.
>>>>>
>>>>> Thanks for your attention.
>>>>>
>>>>> Curt
>>>>>
>>>>>
>>>>> -----Original Message-----
>>>>> From: Neerenberg, Aaron [mailto:[hidden email]]
>>>>> Sent: Friday, July 30, 2010 9:46 AM
>>>>> To: [hidden email]
>>>>> Subject: RE: RE: SCM change when there is no change
>>>>>
>>>>> Not really. �I was wondering if the polling output might have just not
>>>>> saved into the change set correctly, but that doesn't seem to be the
>>>>> case.
>>>>>
>>>>> Regards,
>>>>> Aaron
>>>>>
>>>>> -----Original Message-----
>>>>> From: Curt Patrick [mailto:[hidden email]]
>>>>> Sent: Friday, July 30, 2010 9:39 AM
>>>>> To: '[hidden email]'
>>>>> Subject: RE: SCM change when there is no change
>>>>>
>>>>> Here is the entire log in all its glory. �Does it give you some clue
>>>>> that I'm missing?:
>>>>>
>>>>> Started on Jul 30, 2010 9:37:39 AM
>>>>> Looking for changes...
>>>>> Using remote perforce client:
>>>>> hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>>> [OE-HUBS-Oasis-Java-dev-build] $ p4 workspace -o
>>>>> hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>>> Saving modified client hudson-OE-HUBS-Oasis-Java-dev-build--1273170058
>>>>> [OE-HUBS-Oasis-Java-dev-build] $ p4 -s client -i Last sync'd change
>>>>> was
>>>>> 398604 [OE-HUBS-Oasis-Java-dev-build] $ p4 changes -m 2
>>>>> //hudson-OE-HUBS-Oasis-Java-dev-build--1273170058/...
>>>>> Latest submitted change selected by workspace is 398604 Assuming that
>>>>> the workspace definition has not changed.
>>>>> Done. Took 0.55 sec
>>>>> No changes
>>>>>
>>>>> -----Original Message-----
>>>>> From: Neerenberg, Aaron [mailto:[hidden email]]
>>>>> Sent: Friday, July 30, 2010 9:35 AM
>>>>> To: [hidden email]
>>>>> Subject: RE: SCM change when there is no change
>>>>>
>>>>> Does anything show in the job's Polling log?
>>>>>
>>>>> -----Original Message-----
>>>>> From: Curt Patrick [mailto:[hidden email]]
>>>>> Sent: Friday, July 30, 2010 9:33 AM
>>>>> To: '[hidden email]'
>>>>> Subject: SCM change when there is no change
>>>>>
>>>>> Hello,
>>>>>
>>>>> Can anyone explain what it means when a job says it is "Started by an
>>>>> SCM change" but there is nothing reported in the "Summary of Changes"
>>>>> section? �This happens fairly often and puzzles me. �BTW, I'm using
>>>>> Perforce.
>>>>>
>>>>> Curt
>>>>>
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>>
>>>>
>>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: SCM change when there is no change

Alan Harder-2
maybe someone could cron a little job to copy/backup the polling log
file on the same interval as your jobs do polling, so when this happens
you could look back and see what it gave for the reason it triggered a
build..

    - Alan


Richard Bywater wrote:

> I'm on the same page as Curt re the polling logs. Yes I can see the
> "current" polling log but once the system has polled again, the log is
> overwritten and therefore one can't go back and try and troubleshoot
> the multiple builds.
>
> Richard.
>
> On 9/28/10, Curt Patrick <[hidden email]> wrote:
>  
>> When you look at the page for the build in P4V it says "No changes from the
>> last build".  And below that it says "Started by an SCM change".  That sure
>> doesn't make much sense.  I asked what that meant some time ago and got no
>> useful feedback.
>>
>> Richard may be saying something different than I am, but my problem with the
>> polling log is that, every time, by the time I realize there is a problem
>> and look at the log it has already polled again.  So I can never catch a
>> polling problem "in the act".
>>
>> Curt
>>
>> -----Original Message-----
>> From: Kedar Mhaswade [mailto:[hidden email]]
>> Sent: Monday, September 27, 2010 3:38 PM
>> To: [hidden email]
>> Subject: Re: SCM change when there is no change
>>
>> On 09/27/2010 02:52 PM, Richard Bywater wrote:
>>    
>>> Just to throw another SCM type into the mix, I'm also seeing this on
>>> Clearcase - unfortunately because the polling logs aren't kept (I
>>> don't think?) then I can't tell why it kicked off multiple builds.
>>>
>>> Perhaps the polling log mechanism could be changed to keep the logs
>>> for a period to help solve issues like this?
>>>
>>>      
>> I think the polling logs are (supposed to be) kept in the file
>> <Hudson-Home>/jobs/<job-name>/scm-polling.log when SCM trigger
>> is configured. (I find that at least in the case of subversion).
>>
>> Are you saying you don't find that file on disk?
>>
>> Also, interestingly, the log output that Thomas sent says:
>> "no change for ...", so it appears that Hudson knows there is no change
>> in SCM
>> and still, a new build is scheduled and run.
>>
>>    
>>
>  

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: SCM change when there is no change

Junichi Kimura
In reply to this post by Curt Patrick
On Subversion I have seen this behavior when the master and slave's
clock is off by more than 30 seconds.  I don't recall which direction.

2010/7/30 Curt Patrick <[hidden email]>:

> Hello,
>
> Can anyone explain what it means when a job says it is "Started by an SCM change" but there is nothing reported in the "Summary of Changes" section?  This happens fairly often and puzzles me.  BTW, I'm using Perforce.
>
> Curt
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>



--
Junichi "J" Kimura
[hidden email]

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: SCM change when there is no change

Steele, Richard-2
In reply to this post by Alan Harder-2
I noticed in the changelog for the latest version of Hudson (1.3.78):

If a polling initiated a build, capture its log to the build.

Would that help?

I've also seen this problem using ClearCase.  I've determined it happens when the polling interval comes around and no slave with the job's view is available to do the lshistory needed to check for changes, so by default it just kicks off a build.  It's very annoying but I've come to live with it.

Rich

On Mon, Sep 27, 2010 at 7:52 PM, Alan Harder <[hidden email]> wrote:
maybe someone could cron a little job to copy/backup the polling log file on the same interval as your jobs do polling, so when this happens you could look back and see what it gave for the reason it triggered a build..

  - Alan



Richard Bywater wrote:
I'm on the same page as Curt re the polling logs. Yes I can see the
"current" polling log but once the system has polled again, the log is
overwritten and therefore one can't go back and try and troubleshoot
the multiple builds.

Richard.

On 9/28/10, Curt Patrick <[hidden email]> wrote:
 
When you look at the page for the build in P4V it says "No changes from the
last build".  And below that it says "Started by an SCM change".  That sure
doesn't make much sense.  I asked what that meant some time ago and got no
useful feedback.

Richard may be saying something different than I am, but my problem with the
polling log is that, every time, by the time I realize there is a problem
and look at the log it has already polled again.  So I can never catch a
polling problem "in the act".

Curt

-----Original Message-----
From: Kedar Mhaswade [mailto:[hidden email]]
Sent: Monday, September 27, 2010 3:38 PM
To: [hidden email]
Subject: Re: SCM change when there is no change

On 09/27/2010 02:52 PM, Richard Bywater wrote:
   
Just to throw another SCM type into the mix, I'm also seeing this on
Clearcase - unfortunately because the polling logs aren't kept (I
don't think?) then I can't tell why it kicked off multiple builds.

Perhaps the polling log mechanism could be changed to keep the logs
for a period to help solve issues like this?

     
I think the polling logs are (supposed to be) kept in the file
<Hudson-Home>/jobs/<job-name>/scm-polling.log when SCM trigger
is configured. (I find that at least in the case of subversion).

Are you saying you don't find that file on disk?

Also, interestingly, the log output that Thomas sent says:
"no change for ...", so it appears that Hudson knows there is no change
in SCM
and still, a new build is scheduled and run.

   
 

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]


Reply | Threaded
Open this post in threaded view
|

Re: Multiple branches best practice ?

majje
In reply to this post by Dean Yu
Hi

In the current version of Jenkins you "can" only use one job against one branch of your source. Otherwise the results graphs over time will present a mix of results from several branches, as Dean stated. That's of course not desirable. The idea of copying the job, either manually, with scripts or behind the scene in Jenkins sounds as rough to me.

My idea is to give the job an option to store and present information from different parameter configurations separately. A problem to solve would probably be how to interface this to external consumers of job information (such as views). They need to know about different "versions" of the same job, or the one job with several configurations have to be presented as several jobs to the consumer.

I think it resembles a kind of merge of the job template plugin and the matrix job functionality.

Kind regards

/Magnus
Reply | Threaded
Open this post in threaded view
|

Re: Multiple branches best practice ?

sagetutoring
This post has NOT been accepted by the mailing list yet.
This post was updated on .
In reply to this post by Alex Barna
Why don't you use remote management techniques. in this way you could also make your project more cost effective.
12