Does anybody have an idea how that "-1" is getting generated and where
they might start to look to track down the problem. I can see from
the source code that the string is created with
name += PartitionUtils.SEP + PeriodType.getCalendar().fromIso(
period.getStartDate() ).getYear();
For some reason that getYear() is returning -1 which means they must
have some dodgy period as input. Jason has pointed out that for
analytics, the period might not be coming from the period table, but
rather calculated from incidentdate or enrollmentdate or the like.
If anyone has some helpful advice to give Tony in tracking this down
would be appreciated. As background, I understand this has happened
after an upgrade to postgres 9.5 but this doesn't look like a database
version problem to me.
I am not too sure of the dhis2 version but perhaps Tony can elaborate.
Thanks Bob.
I am currently running dhis 2.22 version 22086 and upgraded the Postgres from version 9.3 to version 9.5. The analytics was successful in the previous Postgres versions 9.3 and the dhis version 2.21.
Strange enough, the tracker instance is having no issue with running analytics.It starts fine and successfully ends, without any errors.This failure in analytics, is happening to just one instances (dhims) which holds both aggregated values and individual events.
Does anybody have an idea how that “-1” is getting generated and where
they might start to look to track down the problem. I can see from
the source code that the string is created with
name += PartitionUtils.SEP + PeriodType.getCalendar().fromIso(
period.getStartDate() ).getYear();
For some reason that getYear() is returning -1 which means they must
have some dodgy period as input. Jason has pointed out that for
analytics, the period might not be coming from the period table, but
rather calculated from incidentdate or enrollmentdate or the like.
If anyone has some helpful advice to give Tony in tracking this down
would be appreciated. As background, I understand this has happened
after an upgrade to postgres 9.5 but this doesn’t look like a database
version problem to me.
I am not too sure of the dhis2 version but perhaps Tony can elaborate.