[ZODB-Dev] RE: [Zope-CMF] Big CMF sites / Storage

[Itamar Shtull-Trauring]

Barry A. Warsaw wrote:


> I'm fairly well convinced that the write performance of
> BerkeleyStorage is limited by BerkeleyDB's inability to handle big
> blob writes well.


I've had this impression too when comparing Btrees+ZODB+FileStorage to 
bsddb3 write perfomance. The issue seems to be especially bad in 
transactional mode - in concurrent mode bsddb3 runs faster, though not 
faster than FileStorage (because you're not writing the data to a 
transactional log in *addition* to the database.) Note that I was using 
plain bsddb3 here, not BerkeleyStorage.