On 07/23/2014 02:13 PM, julie dubois wrote:
I'm using a MySQL database !!!
But may be such database-lock can appear because of my universe.ini
configuration file.
Here it's the lines concerning the database :
database_engine_option_pool_size = 10
database_engine_option_max_overflow = 20
database_engine_option_pool_recycle = -1
database_engine_option_strategy = threadlocal
Can one of these parameters be at the origin of my problem ?
we are also using MySQL, but we use the suggested setting:
database_engine_option_pool_recycle = 7200
though, i don't think this is related to your problem. Or did you get
the error: "MySQL server has gone away" in the galaxy log file?
If yes, when I'll restart Galaxy, the queued jobs will be in a staying
state with the capacity to become in a running state ?
Or they will be in error state and I should restart the process as new jobs ?
I don't know, I don't have enough experience with queued workflow jobs
and what is happening with them during a re-start. I am sorry, I can't
be of much further help. Maybe someone else on the list?
Hans-Rudolf
Thanks.
Julie
2014-07-23 12:55 GMT+02:00 Hans-Rudolf Hotz <hrh(a)fmi.ch>:
>
>
> On 07/23/2014 12:05 PM, julie dubois wrote:
>>
>> Hi Hans-Rudolf,
>> Thanks for your answer.
>> I've launched lot of jobs since monday. And some jobs were finished
>> and some jobs never started.
>> In "Manage jobs" I can see jobs queued because they need inputs which
>> are queued, so I think it's normal.
>>
>> But other jobs are described like this :
>>
>> 9554 user_login 0 hours ago samtools_flagstat new 23199
>> ok, 23243 ok None None None
>> The two necessary inputs for this job (23199 and 2343) are ok and the
>> job doesn't run.
>>
>>
>> Another mysterious line is :
>>
>> 9561 user_login 0 hours ago fastqc new Unable to
>> determine inputs None None None
>>
>> After check this job in my history, I'm sure that the input of this
>> job is green and really available. Why can galaxy not determine inputs
>> ?
>> Finally, in "Manage Jobs" there are only jobs queued. No jobs running.
>>
>> Any idea ?
>
>
> I have one suspicion:
>
> which database are you using?
> If you are just using the built in SQlite you might have run in a
> database-lock? Because you have submitted too many jobs at once?
>
> In this case, I guess, the only way to recover is to re-start...
>
> Hans-Rudolf
>
>
>
>> Thanks
>> Julie
>>
>> 2014-07-23 11:35 GMT+02:00 Hans-Rudolf Hotz <hrh(a)fmi.ch>:
>>>
>>> Hi Julie
>>>
>>> Are there any jobs running at all? What do you get when you click on
>>> "Manage jobs" in the admin menu? Maybe you have some old jobs
'running'
>>> which are now blocking your queue?
>>>
>>> Hans-Rudolf
>>>
>>>
>>>
>>> On 07/23/2014 11:14 AM, julie dubois wrote:
>>>>
>>>>
>>>> Hi all,
>>>> I've seen several posts for jobs which stay queued and never start
>>>> with an installation of galaxy on a cluster.
>>>> But our installation is just a local instance of galaxy in 1 PC and I
>>>> have more than 100 jobs which stay queued and never start.
>>>>
>>>> How can I debug this problem (my version of galaxy is the latest
>>>> 2june2014).
>>>>
>>>> And is it possible to run those jobs without restart galaxy because it
>>>> would be very long to rerun each job manually after reloading galaxy.
>>>>
>>>> Thanks.
>>>>
>>>> Julie
>>>> ___________________________________________________________
>>>> 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/
>>>>
>>>
>