Hi, the initial query performance after a start/restart or after long activity may be dominated by disk performance. When the database is started and data is not in RAM, ArangoDB will need to read everything from disk. The same is true after a long activity if the operating system has decided to page out some of the data to make room for other applications.
I wouldn't be surprised that you will see high disk I/O or disk wait times when you run top (or something similar) while executing the initial queries. Can you try that? How big is that collection on disk approximately? It may be bigger than the 10 MB dump if the datafiles still contain some outdated/already deleted documents. Best regards Jan Am Donnerstag, 16. November 2017 07:52:37 UTC+1 schrieb Andreas Jung: > > Hi there, > > my dev box requires for the first query in the morning extremly long. > First query on a table about 7.000 simple items is > > > for doc in kataster > RETURN DISTINCT doc.gebaeude > > 23.3343360424 seconds > > Subsequent queries are fast > > for doc in kataster > RETURN DISTINCT doc.gebaeude > > 0.245573043823 seconds > > In same case it took up to 90 seconds for the first query. > > ok, data might not be in memory but it should be fast reading the 7.000 > rows > in memory. The overall dump of the related collection is less than 10 MB... > any idea? > > Andreas > > -- You received this message because you are subscribed to the Google Groups "ArangoDB" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
