Autoroute Retry Interval too Long

[mukoya]

The option to specify retry interval has its units in hours.

This is very long. Changing it to minutes would be very useful.

Furthermore, I see there is an option in registry to change maximum retries. That is great. however, an option in the UI would be easier to use and making it configurable per rule would be even better.

Thanks.

[pacsone]

The database job scheduler of PacsOne Server uses the 24-hour clock intervals to schedule all jobs including the retries, and that's why the retry intervals have the resolution of hours instead of minutes or seconds. If you do not want to delay hours for the retries, then don't specify any retry interval which will make PacsOne use the default of retrying the failed jobs immediately after they fail.

It may be possible to make the number of retries configurable by storing in a database table instead of Registry, but this could be a performance drag when there are many failed jobs, e.g., network link down, etc, as each failed job implies reading from that database table in order to check if the number of retry attempts has exceeded the configured maximum (vs checking against a constant read only once from the Registry setting during startup).

[mukoya]

Thanks pacsone for your explanation.

My thinking was that for example if a node is down at a given time, it is less likely to be available in another one second than in, say, 5 minutes. Yet pacsone will continuously keep trying. I thought we might save resources by spreading the retries a little.

It seems the jobs in Pacsone are instance by instance? As you have explained, this could be an issue for cross-sectional imaging modalities with large numbers of images. However, if this were to be at series or study level, the load would be much less. Coupled with reasonable spacing and occasional user intervention e.g, temporarily disabling routing rule for prolonged downtime of destination node, this could work.

Not sure about the implications of limiting jobs to study or series level. Probably some delays in forwarding while awaiting arrival of all/many images to prevent too many associations among others?

The number of retries seems to be set at 3. Can this number be increased? The setting in registry referred to in the user manual does not seem to exist on my installation (6.4.5, Win Server 2012R2)

[pacsone]

It seems the jobs in Pacsone are instance by instance?

No, the Job Levels can be either INSTANCE, SERIES, STUDY or PATIENT, i.e., you can forward by instances, series, studies or patients.

The number of retries seems to be set at 3. Can this number be increased? The setting in registry referred to in the user manual does not seem to exist on my installation (6.4.5, Win Server 2012R2)

The default retry attempts is 3, and you can change it by adding/setting that Registry value then restart PacsOne Server.

[mukoya]

Thanks.

When you say "Job Levels can be either INSTANCE, SERIES, STUDY or PATIENT", Is there a place to set this for auto route jobs or is it left to pacsone to decide?

I will create the registry value and see how it works out.

[pacsone]

When you configure an Automatic Routing rule, there are 2 options named "Wait N Minutes to Forward instances of the entire series" and "Wait N Minutes to Forward instances of the entire study", and if you enable one of them PacsOne will forward the entire series (SERIES level) or study (STUDY level) via one job instead of the default one job per instance.

[mukoya]

I guess then setting rule to do series/study level forward with immediate retry should might be a little lighter than instance level.

Thanks.