Date: Sun, 7 Nov 1999 23:05:49 -0500 (EST) From: Mike Squires <msquires@sir-alan.chem.indiana.edu> To: ports@freebsd.org (FreeBSD ports) Subject: samba@samba.org Message-ID: <199911080405.XAA46023@sir-alan.chem.indiana.edu>
next in thread | raw e-mail | index | archive | help
samba 2.0.6pre3 compiled with patches -aa through -ah and -ba modified *starting line is 1134, changed from 1064) and without -bb (not required) runs about 5X faster on a P5/90 under 3.3-RELEASE. Behavior is now predictable (i.e., transfer rate for multiple tests is approximately the same for transfers of the same size. I used socket options = TCP_NODELAY SO_SNDBUF=65535 SO_RCVBUF=65535 and got max speed of about 200K/sec for 2.0.5a with patches aa-ah, 900K for 2.0.6pre3 with a 13MB file, 650K with 131MB - about what I get with mars_nwe. Hardware was a Micronics M54PE (single P90), AH2944UW with 4 2GB Fujitsu narrow diff drives, Intel Pro100B, 32MB RAM. I have not yet had success with any of the 2.x versions and SMP on a dual PPro system using exactly the same filesystem (and the same binaries for 2.0.5a). This includes two completely different sets of MB/CPU/memory/ controllers. I currently would be very interested to know if anyone has samba 2.x running reliably on an SMP system using PPro or PII/PIII CPUs. MIke Squires mikes@indiana.edu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-ports" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199911080405.XAA46023>