Till Vollmer
2009-Mar-30 09:59 UTC
[Backgroundrb-devel] Severe Bug with daylight saving time?
Hi,
we have encountered a severe bug in backgroundrb last weekend.
On Sunday at 20:00 (EST) backgroundrb launched infinite workers of
notification_worker and our server was killed because of memory
consumption.
:backgroundrb:
:port: 11006
:ip: 0.0.0.0
:persistent_disabled: true # turn this off if your application
doesn''t use backgroundrb''s persistent/enqueued tasks system
:persistent_delay: 10 # the time (seconds) between each time
backgroundrb checks the database for enqueued tasks
:schedules:
:notification_worker:
:do_it:
:trigger_args: 0 30 */4 * * * *
We think that it might be related to daylight saving time.
Any hints appreciated.
Regards
Till
MindMeister - MeisterLabs GmbH
Till Vollmer
Managing Director
Tel: +49 (0)89 1213 5359
Mob: + 49 (0)160 718 7403
Fax: +49 (0)89 1892 1347
Yahoo ID: till_vollmer
Skype: till_vollmer
www.mindmeister.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://rubyforge.org/pipermail/backgroundrb-devel/attachments/20090330/a158de67/attachment.html>
On Mon, Mar 30, 2009 at 3:29 PM, Till Vollmer <till at mindmeister.com> wrote:> Hi, > we have encountered a severe bug in backgroundrb last weekend. > On Sunday at 20:00 (EST) backgroundrb launched infinite workers of > notification_worker and our server was killed because of memory consumption. > :backgroundrb: > ??:port: 11006 > ??:ip: 0.0.0.0 > ??:persistent_disabled: true # turn this off if your application doesn''t use > backgroundrb''s persistent/enqueued tasks system > ??:persistent_delay: 10 # the time (seconds) between each time backgroundrb > checks the database for enqueued tasks > :schedules: > ??:notification_worker: > ?? ?:do_it: > ?? ? ?:trigger_args: 0 30 */4 * * * * > We think that it might be related to daylight saving time. > Any hints?appreciated.Were you using reload_on_schedule ?? But yes, the bug do seem like triggered by DST, I will be pushing out an update that will contain the fix. BTW, were you able to recover the worker?
Till Vollmer
2009-Mar-30 12:46 UTC
[Backgroundrb-devel] Severe Bug with daylight saving time?
Hi,
The worker looks like that:
class NotificationWorker < BackgrounDRb::MetaWorker
set_worker_name :notification_worker
reload_on_schedule true
def create(args = nil)
# time argument is in seconds
end
def do_it
logger.info(''NotificationWorker do it start'')
begin
Idea.notify_users
rescue Exception => e
logger.error("Idea.notify_users Exception: #{e.to_s}
#{e.backtrace.join(''\\n'')}")
end
logger.info(''NotificationWorker do it end'')
end
end
Recover - we removed the scheduling and trigger only daily right now.
Not sure if it would work today (after the day that has only 23
hours). but as it is a production system we did not try it.
Regards
Till
MindMeister - MeisterLabs GmbH
Till Vollmer
Managing Director
Tel: +49 (0)89 1213 5359
Mob: + 49 (0)160 718 7403
Fax: +49 (0)89 1892 1347
Yahoo ID: till_vollmer
Skype: till_vollmer
www.mindmeister.com
Am 30.03.2009 um 14:31 schrieb hemant:
> On Mon, Mar 30, 2009 at 3:29 PM, Till Vollmer <till at
mindmeister.com>
> wrote:
>> Hi,
>> we have encountered a severe bug in backgroundrb last weekend.
>> On Sunday at 20:00 (EST) backgroundrb launched infinite workers of
>> notification_worker and our server was killed because of memory
>> consumption.
>> :backgroundrb:
>> :port: 11006
>> :ip: 0.0.0.0
>> :persistent_disabled: true # turn this off if your application
>> doesn''t use
>> backgroundrb''s persistent/enqueued tasks system
>> :persistent_delay: 10 # the time (seconds) between each time
>> backgroundrb
>> checks the database for enqueued tasks
>> :schedules:
>> :notification_worker:
>> :do_it:
>> :trigger_args: 0 30 */4 * * * *
>> We think that it might be related to daylight saving time.
>> Any hints appreciated.
>
> Were you using reload_on_schedule ?? But yes, the bug do seem like
> triggered by DST, I will be pushing out an update that will contain
> the fix.
>
> BTW, were you able to recover the worker?
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://rubyforge.org/pipermail/backgroundrb-devel/attachments/20090330/ee6863a7/attachment-0001.html>
Adam Williams
2009-Mar-30 13:47 UTC
[Backgroundrb-devel] Severe Bug with daylight saving time?
I have suffered the same way, and also believe it to be a time change issue. We spent a good deal of time discovering, after the time change a few weeks ago, that the schedule ''0 0 0 * * * *'' would no long work at all. We are now running on ''23 59 59 * * * *'', and that seems to be fine. On Mar 30, 2009, at 5:59 AM, Till Vollmer wrote:> Hi, > > we have encountered a severe bug in backgroundrb last weekend. > > On Sunday at 20:00 (EST) backgroundrb launched infinite workers of > notification_worker and our server was killed because of memory > consumption. > > :backgroundrb: > :port: 11006 > :ip: 0.0.0.0 > :persistent_disabled: true # turn this off if your application > doesn''t use backgroundrb''s persistent/enqueued tasks system > :persistent_delay: 10 # the time (seconds) between each time > backgroundrb checks the database for enqueued tasks > > :schedules: > :notification_worker: > :do_it: > :trigger_args: 0 30 */4 * * * * > > We think that it might be related to daylight saving time. > > Any hints appreciated. > > Regards > Till > > MindMeister - MeisterLabs GmbH > Till Vollmer > Managing Director > Tel: +49 (0)89 1213 5359 > Mob: + 49 (0)160 718 7403 > Fax: +49 (0)89 1892 1347 > Yahoo ID: till_vollmer > Skype: till_vollmer > www.mindmeister.com > > > _______________________________________________ > Backgroundrb-devel mailing list > Backgroundrb-devel at rubyforge.org > http://rubyforge.org/mailman/listinfo/backgroundrb-devel
Till Vollmer
2009-Apr-06 11:59 UTC
[Backgroundrb-devel] Severe Bug with daylight saving time?
Hi, Did you have time to fix this? Is it mirrored to svn already? Regards Till MindMeister - MeisterLabs GmbH Till Vollmer Managing Director Tel: +49 (0)89 1213 5359 Mob: + 49 (0)160 718 7403 Fax: +49 (0)89 1892 1347 Yahoo ID: till_vollmer Skype: till_vollmer www.mindmeister.com Am 30.03.2009 um 14:31 schrieb hemant:> On Mon, Mar 30, 2009 at 3:29 PM, Till Vollmer <till at mindmeister.com> > wrote: >> Hi, >> we have encountered a severe bug in backgroundrb last weekend. >> On Sunday at 20:00 (EST) backgroundrb launched infinite workers of >> notification_worker and our server was killed because of memory >> consumption. >> :backgroundrb: >> :port: 11006 >> :ip: 0.0.0.0 >> :persistent_disabled: true # turn this off if your application >> doesn''t use >> backgroundrb''s persistent/enqueued tasks system >> :persistent_delay: 10 # the time (seconds) between each time >> backgroundrb >> checks the database for enqueued tasks >> :schedules: >> :notification_worker: >> :do_it: >> :trigger_args: 0 30 */4 * * * * >> We think that it might be related to daylight saving time. >> Any hints appreciated. > > Were you using reload_on_schedule ?? But yes, the bug do seem like > triggered by DST, I will be pushing out an update that will contain > the fix. > > BTW, were you able to recover the worker?-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://rubyforge.org/pipermail/backgroundrb-devel/attachments/20090406/58b4adcd/attachment.html>