Issue #2297: several times a day queries take several minutes

we have queries like (see the database that we handed over to you for other issues)

select distinct o from ObjectNode o join o.properties p1 join o.properties p2 where o.classIdentifier = "(OP)" and (o.linkedObjects.classIdentifier = "(TC)" and  o.linkedObjects.objectNameUpper = "MILLING-EL" and o.linkedObjects.type =1) and (o.properties.name = "feedbackState" and o.properties.doubleValue != 13) and ((p1.name = "plannedState" and p1.doubleValue = 4 and p1.state = 4) or ((p1.name = "startableState" and p1.doubleValue = 0) and (p2.name = "plannedState" and p2.doubleValue = 4)))

that normally take around 10 seconds but sometimes several times a day take around 3-10 minutes. Our system has a load of around 400 concurrent webSocket connections that normally only query (without updates, like the above example)

Can you monitor the server machine and identify unusual activity at the server when that happens? e.g. other operations against ObjectDB, other server activity. Hard drive / memory usage?

Is it related to issue #2257 that is still open?

Suspected configuration:

• <temp ... threshold="64mb" />
  Consider increasing the threshold to avoid using temporary files on query execution.

• <recovery ... max="128mb" />
  Consider decreasing the maximum size to force updating the database file more often to spread massive write operations.

we do not see a link to the open issue as the database seems to be ok.

we'll change the options accordingly

OK. Please report if these configuration changes have any effect on this issue.