Hi. > Hi! >> When extracting large zip packages (~1.6Gb), Java heap space error is >> thrown. Currently used attribute -Xmx500M isn't enough. Memory usage >> rises >> gradually during extraction (slowly, but surely). We have a need to >> handle >> packages up to 2Gb size. >> > Do this zip file contains many entries? > If this is the case then we are out of luck currently. VFS reads the > whole directory structure in memory. > I dont see what we can do to avoid this without introducing a HUGE > performance impact.
Yes, the zip files contain a lot of entries. However, we're not experiencing problems when we extract say 1 big zip. The problems occur later, when we have extracted like 5 of them. It seems like the VFS doesn't free the used memory, even when the object used is out of scope. Also, we've tried to clear cache with getFilesCache().clear(usedFileObject.getFileSystem()). Even this doesn't seem to help. Do you see this as a potential VFS problem? > If this is not the case it would be nice if you could start up a > profiler to see what uses so much memory, maybe its again a > java.util.zip limitation? Did you try some simple code using > java.util.zip directly? Could you point me to a working profiler? I have tried to found some for Eclipse 3.1.2, but I'm out of luck. Thanks a lot for your help. BR, -hrjk --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
