Inline...
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 12/29/2013 5:11 PM, Pierre Smits wrote:
Adrian, Rupert,
The renamed job pool was only an example. I haven't renamed any job in the
setup as I understood that there might be predicaments with dependencies in
a running environment.
Based on the feedback given by both of you, I came up with following
configuration:
On any front end (user oriented) server I keep the settings regarding the
original thread-pool as-is, but set the poll-enabled attribute to 'false'.
On any back end server I also keep the settings regarding the thread-pool
as-is.
This will ensure that existing (in seed, seed-initial and hardcoded) job
definition keeps working as-is.
Additionally I create on the (new) front end server a new thread-pool like
this:
<thread-pool send-to-pool="pool"
^^^Indicates jobs originating on this server will be sent to the job
pool called "pool".
purge-job-days="4"
failed-retry-min="3"
ttl="120000"
jobs="100"
min-threads="2"
max-threads="5"
poll-enabled="true"
^^^Indicates this server will poll (query) the database for new jobs to run.
poll-db-millis="30000">
<run-from-pool name="FrontEnd01pool"/>
^^^Indicates this server will run jobs sent to the "FrontEnd01pool"
pool. Jobs sent to the pool named "pool" will need to be run on a
different server.
</thread-pool>
This will enable us to be flexible with regards to where new jobs (not from
seed, seed-initial and hardcoded) are initiated and where they are
executed. Or am I mistaken?
With regards to those with pool hardcoded, is it correctable or is it
required?
The hard-coded pool name should be replaced with code that gets the
send-to-pool setting from the service engine.
I don't see how we can "fix" the seed/demo data issue - the included
data must make some assumptions about the default setup.
Regards,
Pierre Smits
*ORRTIZ.COM <http://www.orrtiz.com>*
Services & Solutions for Cloud-
Based Manufacturing, Professional
Services and Retail & Trade
http://www.orrtiz.com
On Sun, Dec 29, 2013 at 2:36 PM, Adrian Crum <
[email protected]> wrote:
One more thing - I noticed you are renaming the job pools. That is a bad
idea. I've noticed some code has the "pool" job pool hard-coded, and the
seed/demo data expects there to be a "pool" job pool. (Renaming pools is
conceptually correct, but not all developers have followed that design
pattern.) So, there is a chance some things won't work when you rename the
pools.
So, to implement what you described, just leave everything as-is and only
change the poll-enabled="true" setting as I mentioned before.
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 12/29/2013 8:28 AM, Adrian Crum wrote:
No, that is not correct. Set poll-enabled="true" only on the server that
will be servicing the jobs, all others should be set to "false".
The schema contains documentation on how to set up the Job Scheduler. If
there is missing information, or if the information is not clear, then
please let me know and I will update it.
Also be aware that each sever must have a unique name - specified in the
unique.instanceId property in general.properties.
Adrian Crum
Sandglass Software
www.sandglass-software.com
On 12/29/2013 8:16 AM, Rupert Howell wrote:
Pierre. This is correct.
All front end initiated asynchronous services will be executed by the
backend server and the backend server will run its own in this
configuration.
Rupert Howell
Provolve Ltd.
Technopole
Kingston Crescent
Portsmouth
PO2 8FA
07909 685308
http://www.provolve.com
On 29 Dec 2013, at 13:01, Pierre Smits <[email protected]> wrote:
Hi Rupert,
Thanks for the info.
So to understand the setting in serviceengine.xml correctly I would have
the following on the backend spoke:
<thread-pool send-to-pool="*BackEndPool01*>"
purge-job-days="4"
failed-retry-min="3"
ttl="120000"
jobs="100"
min-threads="2"
max-threads="5"
poll-enabled="true"
poll-db-millis="30000">
<run-from-pool name="*BackEndPool01*"/>
</thread-pool>
And on the frontend spoke I would have:
<thread-pool send-to-pool="*BackEndPool01*"
purge-job-days="4"
failed-retry-min="3"
ttl="120000"
jobs="100"
min-threads="2"
max-threads="5"
poll-enabled="true"
poll-db-millis="30000">
<run-from-pool name="*FrontEndPool01*"/>
</thread-pool>
And set the services on any spoke to point to where I want to have it
executed, eg:
<JobSandbox jobId="8200" jobName="Clear EntitySyncRemove Info"
runTime="2000-01-01
00:00:00.000" serviceName="cleanSyncRemoveInfo" poolId="*BackEndPool01*"
runAsUser="system" tempExprId="MIDNIGHT_DAILY" maxRecurrenceCount="-1"/>
Is this assumed correctly?
Regards,
Pierre Smits
*ORRTIZ.COM <http://www.orrtiz.com>*
Services & Solutions for Cloud-
Based Manufacturing, Professional
Services and Retail & Trade
http://www.orrtiz.com
