Did you put a hash index on the trackid attribute? If not, a full collection scan is started, which means the system has to look at every single document and has thus to load the data into RAM. If the mass storage is slow, or very little main memory is available for the memory mapped files, then it is only natural that performance degrades.
If not even the primary indexes fit in the memory, then it's going to be worse, because the OS will start to swap data between RAM and mass storage. If the track-ids are unique and your foremost way to fetch documents, then I suggest to use use the _key attribute instead. It comes with a hash index on it and you can save yourself another hash index on the track-id. -- 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.
