On Wed, 17 Nov 2004 at 8:06:22 -0600, C. Bensend wrote: > > > Until it is added, you'll have to use clamscan. Note though, that if you > > scan many files at once (with one launching of clamscan), performance > > increasing thanks to using clamdscan instead of clamscan - is > > negligible (because the program and the database are loaded only one > > time). > > Yes, I just noticed that last night - until then, my performance tests > had been on small amounts of files, where the difference is quite large > (12s -vs- 0.5s, etc). > > However, when I opened up clamscan and clamdscan on a real directory > (114,000 files), the totals were like 38m (clamscan) -vs- 34m > (clamdscan).
Such big (38 min. vs. 34 min.) difference while scanning so many files isn't likely due to difference between clamscan and clamdscan per se. It can be seconds, not minutes. I suppose you have run clamscan first, then clamdscan?... Subsequent accessing files is far more quicker thanks to buffering by the operating system. Here you are a real example. I've just run clamscan on a directory and after the 1st scanning ended, I've run it again. $ time clamscan -ri /usr/share/doc [...] ----------- SCAN SUMMARY ----------- Known viruses: 26866 Scanned directories: 646 Scanned files: 5065 Infected files: 2 Data scanned: 140.39 MB I/O buffer size: 131072 bytes Time: 110.264 sec (1 m 50 s) real 1m50.271s user 1m24.720s sys 0m6.660s $ time clamscan -ri /usr/share/doc [...] ----------- SCAN SUMMARY ----------- Known viruses: 26866 Scanned directories: 646 Scanned files: 5065 Infected files: 2 Data scanned: 140.39 MB I/O buffer size: 131072 bytes Time: 92.476 sec (1 m 32 s) real 1m32.483s user 1m24.890s sys 0m6.050s The difference is 110 s vs. 92 s. One of other reasons can be different system load. One must also remember that clamdscan operation is strictly dependent on clamd.conf. If you, by chance, had there some default options disabled, which are "enabled" for "default" executing of clamscan, then it would make clamdscan's job less laborious, resulting in shorter scanning. But the OS buffering is the most likely reason. -- Tomasz Papszun SysAdm @ TP S.A. Lodz, Poland | And it's only [EMAIL PROTECTED] http://www.lodz.tpsa.pl/iso/ | ones and zeros. [EMAIL PROTECTED] http://www.ClamAV.net/ A GPL virus scanner _______________________________________________ http://lists.clamav.net/cgi-bin/mailman/listinfo/clamav-users
