Hi dan
Is that per dag or per dag bag? Multiprocessing should parallelize dag parsing
so I am very curious. Let me know if I can help out.
Bolke
Sent from my iPhone
> On 3 mei 2016, at 01:47, Dan Davydov wrote:
>
> So a quick update, unfortunately we saw some DAGBag parsing time increases
>
Hey Dan,
Thanks for the update! Please keep us posted.
Cheers,
Chris
On Mon, May 2, 2016 at 4:47 PM, Dan Davydov
wrote:
> So a quick update, unfortunately we saw some DAGBag parsing time increases
> (~10x for some DAGs) on the webservers with the 1.7.1rc3. Because of this I
> will be working o
So a quick update, unfortunately we saw some DAGBag parsing time increases
(~10x for some DAGs) on the webservers with the 1.7.1rc3. Because of this I
will be working on a staging cluster that has a copy of our production
production DAGBag, and is a copy of our production airflow infrastructure,
ju
any further thoughts on this?
On Sat, Apr 30, 2016 at 5:08 PM, harish singh
wrote:
> Is it possible that the sql you're running to get customer ids is not the
> same every time? That's what I (loosely) meant by non-deterministic.
> [response]
> The sql is the same. But it is definitely possible
Hey Neeraj,
I recommend starting with some of the tickets labeled with 'newbie':
https://issues.apache.org/jira/issues/?jql=project%20%3D%20AIRFLOW%20AND%20status%20%3D%20Open%20AND%20labels%20%3D%20newbie
We will be adding more as we migrate issues from GitHub to Apache.
Cheers,
Chris
On Mon
Hello,
I would like to participate and contribute to the project development of
Airflow project. I have 15 years of experience in IT which includes more
than 10 years in design and development of Workflow based application like
PLM TeamCenter, TrackWise etc.
Please let me know, if I can participa
> Op 2 mei 2016, om 13:48 heeft Jeremiah Lowin het volgende
> geschreven:
>
> Well done Bolke and thanks for all the work on this! Looking forward to
> kicking the tires later today.
>
> After last week's conversations I understand why users would want a run_id
> but I'm not totally sold that
Well done Bolke and thanks for all the work on this! Looking forward to
kicking the tires later today.
After last week's conversations I understand why users would want a run_id
but I'm not totally sold that (dag_id, execution_date) is an insufficient
primary key for a DagRun. My reasoning is this
