Cufflinks doesn't stop

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

Cufflinks doesn't stop

Federico Zambelli
Hi,

something strange happens with Cufflinks in our Galaxy server. When a
user deletes a running Cufflinks job in fact the associated Cufflinks
process(es) are not terminated. Apart from unneccessary CPU usage, this
prevents other jobs from starting if the max jobs limit has been already
reached by the user.

This happens only with Cufflinks, Cuffcompare for comparison behaves
just fine.

Best,
F.



--
============================================
Federico Zambelli, Ph.D.
Bioinformatics, Evolution and Comparative Genomics Lab
Dept. of Biosciences
University of Milano - Italy

What can be asserted without proof can be dismissed without proof.
============================================
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Peter Cock
On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
<[hidden email]> wrote:

> Hi,
>
> something strange happens with Cufflinks in our Galaxy server. When a user
> deletes a running Cufflinks job in fact the associated Cufflinks process(es)
> are not terminated. Apart from unneccessary CPU usage, this prevents other
> jobs from starting if the max jobs limit has been already reached by the
> user.
>
> This happens only with Cufflinks, Cuffcompare for comparison behaves just
> fine.
>
> Best,
> F.

What cluster system are you using? Can you try a few other long
running tools for comparison?

I've been meaning to double check if this still happens, but I had seen
this for BLAST+ jobs under SGE (noticeable again when large zombie
jobs are blocking the queue).

Peter
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Federico Zambelli
Hi Peter,

it runs on a single server since this Galaxy instance is for internal
use of our lab only.

Aborting other long running tools like tophat and bowtie works just as
expected. I have also noticed that at least a couple of times not only
the cufflinks process(es) were not terminated but also that one of the
Galaxy handler processes (five of them in our instance) crashed. I
suspect that the two things are related but have no proof for that.

F.

Il 18/02/14 15:24, Peter Cock ha scritto:

> On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
> <[hidden email]> wrote:
>> Hi,
>>
>> something strange happens with Cufflinks in our Galaxy server. When a user
>> deletes a running Cufflinks job in fact the associated Cufflinks process(es)
>> are not terminated. Apart from unneccessary CPU usage, this prevents other
>> jobs from starting if the max jobs limit has been already reached by the
>> user.
>>
>> This happens only with Cufflinks, Cuffcompare for comparison behaves just
>> fine.
>>
>> Best,
>> F.
>
> What cluster system are you using? Can you try a few other long
> running tools for comparison?
>
> I've been meaning to double check if this still happens, but I had seen
> this for BLAST+ jobs under SGE (noticeable again when large zombie
> jobs are blocking the queue).
>
> Peter
>






--
============================================
Federico Zambelli, Ph.D.
Bioinformatics, Evolution and Comparative Genomics Lab
Dept. of Biosciences
University of Milano - Italy

What can be asserted without proof can be dismissed without proof.
============================================
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Björn Grüning-3
In reply to this post by Peter Cock
Hi,

I can also see that behaviour under SGE long running jobs, like blast+.
Screenshot attached.

Cheers,
Bjoern

> On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
> <[hidden email]> wrote:
>> Hi,
>>
>> something strange happens with Cufflinks in our Galaxy server. When a user
>> deletes a running Cufflinks job in fact the associated Cufflinks process(es)
>> are not terminated. Apart from unneccessary CPU usage, this prevents other
>> jobs from starting if the max jobs limit has been already reached by the
>> user.
>>
>> This happens only with Cufflinks, Cuffcompare for comparison behaves just
>> fine.
>>
>> Best,
>> F.
> What cluster system are you using? Can you try a few other long
> running tools for comparison?
>
> I've been meaning to double check if this still happens, but I had seen
> this for BLAST+ jobs under SGE (noticeable again when large zombie
> jobs are blocking the queue).
>
> Peter
> ___________________________________________________________
> Please keep all replies on the list by using "reply all"
> in your mail client.  To manage your subscriptions to this
> and other Galaxy lists, please use the interface at:
>    http://lists.bx.psu.edu/
>
> To search Galaxy mailing lists use the unified search at:
>    http://galaxyproject.org/search/mailinglists/

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/

deleted_but_running.png (38K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Björn Grüning-3
In reply to this post by Peter Cock
Hi,

I have one running at the moment, that job is also not listed in the
jobs monitor under the admin-panel.

Cheers,
Bjoern

> On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
> <[hidden email]> wrote:
>> Hi,
>>
>> something strange happens with Cufflinks in our Galaxy server. When a user
>> deletes a running Cufflinks job in fact the associated Cufflinks process(es)
>> are not terminated. Apart from unneccessary CPU usage, this prevents other
>> jobs from starting if the max jobs limit has been already reached by the
>> user.
>>
>> This happens only with Cufflinks, Cuffcompare for comparison behaves just
>> fine.
>>
>> Best,
>> F.
> What cluster system are you using? Can you try a few other long
> running tools for comparison?
>
> I've been meaning to double check if this still happens, but I had seen
> this for BLAST+ jobs under SGE (noticeable again when large zombie
> jobs are blocking the queue).
>
> Peter
> ___________________________________________________________
> Please keep all replies on the list by using "reply all"
> in your mail client.  To manage your subscriptions to this
> and other Galaxy lists, please use the interface at:
>    http://lists.bx.psu.edu/
>
> To search Galaxy mailing lists use the unified search at:
>    http://galaxyproject.org/search/mailinglists/

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Nate Coraor (nate@bx.psu.edu)
Hi all,

There was a regression introduced in the most recent stable release that was preventing jobs from being stopped, fixed here:


You can pull from the stable branch of galaxy-central to get that changeset. If that doesn't resolve the problem, it sounds like it may be due to the crashing handlers, so we would need to figure out the cause of those crashes.

--nate


On Sat, Feb 22, 2014 at 12:24 PM, Björn Grüning <[hidden email]> wrote:
Hi,

I have one running at the moment, that job is also not listed in the jobs monitor under the admin-panel.

Cheers,
Bjoern


On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
<[hidden email]> wrote:
Hi,

something strange happens with Cufflinks in our Galaxy server. When a user
deletes a running Cufflinks job in fact the associated Cufflinks process(es)
are not terminated. Apart from unneccessary CPU usage, this prevents other
jobs from starting if the max jobs limit has been already reached by the
user.

This happens only with Cufflinks, Cuffcompare for comparison behaves just
fine.

Best,
F.
What cluster system are you using? Can you try a few other long
running tools for comparison?

I've been meaning to double check if this still happens, but I had seen
this for BLAST+ jobs under SGE (noticeable again when large zombie
jobs are blocking the queue).

Peter
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
   http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
   http://galaxyproject.org/search/mailinglists/

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
 http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
 http://galaxyproject.org/search/mailinglists/


___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Björn Grüning-3
Hi Nate,

I'm running

changeset:   12641:cab3db6e1d59

and I can see the same behaviour. The blast job is not listed under jobs
in the admin section and I can't kill it, also not via deleting the dataset.

Anything I can do to track that down?
In my case it's a large blast job with splitting enabled.

Cheers,
Bjoern


> Hi all,
>
> There was a regression introduced in the most recent stable release that
> was preventing jobs from being stopped, fixed here:
>
>
> https://bitbucket.org/galaxy/galaxy-central/commits/1298d3f6aca59825d0eb3d32afd5686c4b1b9294?at=stable
>
> You can pull from the stable branch of galaxy-central to get that
> changeset. If that doesn't resolve the problem, it sounds like it may be
> due to the crashing handlers, so we would need to figure out the cause of
> those crashes.
>
> --nate
>
>
> On Sat, Feb 22, 2014 at 12:24 PM, Björn Grüning
> <[hidden email]>wrote:
>
>> Hi,
>>
>> I have one running at the moment, that job is also not listed in the jobs
>> monitor under the admin-panel.
>>
>> Cheers,
>> Bjoern
>>
>>
>>   On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
>>> <[hidden email]> wrote:
>>>
>>>> Hi,
>>>>
>>>> something strange happens with Cufflinks in our Galaxy server. When a
>>>> user
>>>> deletes a running Cufflinks job in fact the associated Cufflinks
>>>> process(es)
>>>> are not terminated. Apart from unneccessary CPU usage, this prevents
>>>> other
>>>> jobs from starting if the max jobs limit has been already reached by the
>>>> user.
>>>>
>>>> This happens only with Cufflinks, Cuffcompare for comparison behaves just
>>>> fine.
>>>>
>>>> Best,
>>>> F.
>>>>
>>> What cluster system are you using? Can you try a few other long
>>> running tools for comparison?
>>>
>>> I've been meaning to double check if this still happens, but I had seen
>>> this for BLAST+ jobs under SGE (noticeable again when large zombie
>>> jobs are blocking the queue).
>>>
>>> Peter
>>> ___________________________________________________________
>>> Please keep all replies on the list by using "reply all"
>>> in your mail client.  To manage your subscriptions to this
>>> and other Galaxy lists, please use the interface at:
>>>     http://lists.bx.psu.edu/
>>>
>>> To search Galaxy mailing lists use the unified search at:
>>>     http://galaxyproject.org/search/mailinglists/
>>>
>>
>> ___________________________________________________________
>> Please keep all replies on the list by using "reply all"
>> in your mail client.  To manage your subscriptions to this
>> and other Galaxy lists, please use the interface at:
>>   http://lists.bx.psu.edu/
>>
>> To search Galaxy mailing lists use the unified search at:
>>   http://galaxyproject.org/search/mailinglists/
>>
>

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Nate Coraor (nate@bx.psu.edu)
Hi Björn,

Does stopping other, non-split jobs work correctly for you?

--nate


On Wed, Feb 26, 2014 at 7:30 AM, Björn Grüning <[hidden email]> wrote:
Hi Nate,

I'm running

changeset:   12641:cab3db6e1d59

and I can see the same behaviour. The blast job is not listed under jobs in the admin section and I can't kill it, also not via deleting the dataset.

Anything I can do to track that down?
In my case it's a large blast job with splitting enabled.

Cheers,
Bjoern



Hi all,

There was a regression introduced in the most recent stable release that
was preventing jobs from being stopped, fixed here:


https://bitbucket.org/galaxy/galaxy-central/commits/1298d3f6aca59825d0eb3d32afd5686c4b1b9294?at=stable

You can pull from the stable branch of galaxy-central to get that
changeset. If that doesn't resolve the problem, it sounds like it may be
due to the crashing handlers, so we would need to figure out the cause of
those crashes.

--nate


On Sat, Feb 22, 2014 at 12:24 PM, Björn Grüning
<[hidden email]>wrote:

Hi,

I have one running at the moment, that job is also not listed in the jobs
monitor under the admin-panel.

Cheers,
Bjoern


  On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
<[hidden email]> wrote:

Hi,

something strange happens with Cufflinks in our Galaxy server. When a
user
deletes a running Cufflinks job in fact the associated Cufflinks
process(es)
are not terminated. Apart from unneccessary CPU usage, this prevents
other
jobs from starting if the max jobs limit has been already reached by the
user.

This happens only with Cufflinks, Cuffcompare for comparison behaves just
fine.

Best,
F.

What cluster system are you using? Can you try a few other long
running tools for comparison?

I've been meaning to double check if this still happens, but I had seen
this for BLAST+ jobs under SGE (noticeable again when large zombie
jobs are blocking the queue).

Peter
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
    http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
    http://galaxyproject.org/search/mailinglists/


___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/





___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/
Reply | Threaded
Open this post in threaded view
|

Re: Cufflinks doesn't stop

Björn Grüning-3
Hi Nate,

I can kill other jobs and other jobs are also showing up under "Job
list". Not sure if it something related to job splitting, but I only
encountered it with Blast.

Thanks,
Bjoern


> Hi Björn,
>
> Does stopping other, non-split jobs work correctly for you?
>
> --nate
>
>
> On Wed, Feb 26, 2014 at 7:30 AM, Björn Grüning <[hidden email]>wrote:
>
>> Hi Nate,
>>
>> I'm running
>>
>> changeset:   12641:cab3db6e1d59
>>
>> and I can see the same behaviour. The blast job is not listed under jobs
>> in the admin section and I can't kill it, also not via deleting the dataset.
>>
>> Anything I can do to track that down?
>> In my case it's a large blast job with splitting enabled.
>>
>> Cheers,
>> Bjoern
>>
>>
>>
>>   Hi all,
>>>
>>> There was a regression introduced in the most recent stable release that
>>> was preventing jobs from being stopped, fixed here:
>>>
>>>
>>> https://bitbucket.org/galaxy/galaxy-central/commits/
>>> 1298d3f6aca59825d0eb3d32afd5686c4b1b9294?at=stable
>>>
>>> You can pull from the stable branch of galaxy-central to get that
>>> changeset. If that doesn't resolve the problem, it sounds like it may be
>>> due to the crashing handlers, so we would need to figure out the cause of
>>> those crashes.
>>>
>>> --nate
>>>
>>>
>>> On Sat, Feb 22, 2014 at 12:24 PM, Björn Grüning
>>> <[hidden email]>wrote:
>>>
>>>   Hi,
>>>>
>>>> I have one running at the moment, that job is also not listed in the jobs
>>>> monitor under the admin-panel.
>>>>
>>>> Cheers,
>>>> Bjoern
>>>>
>>>>
>>>>    On Tue, Feb 18, 2014 at 2:20 PM, Federico Zambelli
>>>>
>>>>> <[hidden email]> wrote:
>>>>>
>>>>>   Hi,
>>>>>>
>>>>>> something strange happens with Cufflinks in our Galaxy server. When a
>>>>>> user
>>>>>> deletes a running Cufflinks job in fact the associated Cufflinks
>>>>>> process(es)
>>>>>> are not terminated. Apart from unneccessary CPU usage, this prevents
>>>>>> other
>>>>>> jobs from starting if the max jobs limit has been already reached by
>>>>>> the
>>>>>> user.
>>>>>>
>>>>>> This happens only with Cufflinks, Cuffcompare for comparison behaves
>>>>>> just
>>>>>> fine.
>>>>>>
>>>>>> Best,
>>>>>> F.
>>>>>>
>>>>>>   What cluster system are you using? Can you try a few other long
>>>>> running tools for comparison?
>>>>>
>>>>> I've been meaning to double check if this still happens, but I had seen
>>>>> this for BLAST+ jobs under SGE (noticeable again when large zombie
>>>>> jobs are blocking the queue).
>>>>>
>>>>> Peter
>>>>> ___________________________________________________________
>>>>> Please keep all replies on the list by using "reply all"
>>>>> in your mail client.  To manage your subscriptions to this
>>>>> and other Galaxy lists, please use the interface at:
>>>>>      http://lists.bx.psu.edu/
>>>>>
>>>>> To search Galaxy mailing lists use the unified search at:
>>>>>      http://galaxyproject.org/search/mailinglists/
>>>>>
>>>>>
>>>> ___________________________________________________________
>>>> Please keep all replies on the list by using "reply all"
>>>> in your mail client.  To manage your subscriptions to this
>>>> and other Galaxy lists, please use the interface at:
>>>>    http://lists.bx.psu.edu/
>>>>
>>>> To search Galaxy mailing lists use the unified search at:
>>>>    http://galaxyproject.org/search/mailinglists/
>>>>
>>>>
>>>
>>
>

___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client.  To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
  http://lists.bx.psu.edu/

To search Galaxy mailing lists use the unified search at:
  http://galaxyproject.org/search/mailinglists/