Date: Tue, 6 Apr 1999 02:31:57 -0500 (CDT) From: Ken Brownfield <kenb@irridia.com> To: freebsd-questions@freebsd.org Subject: Disk I/O crushes Ethernet I/O (3.0R) Message-ID: <199904060731.CAA10029@asooo.irridia.com>
next in thread | raw e-mail | index | archive | help
Hello, I hope this is the appropriate list for this type of question. Tons of web/mailing list searches haven't turned up anything useful. I'm currently running 3.0R on the following hardware: SuperMicro P6DLE Dual P2/233, 340MB RAM 12GB IDE drive Intel EtherExpress PRO/100B on 10BT network During periods of disk activity (copying a large file, etc.) all network traffic virtually comes to a halt. If I run the cp during an FTP file transfer, my network transfer rate is dropped by about 80%. Other network services suffer the same massive packet deprivation. Any sustained disk access will cause it, and all network transfers (in and out) are effected. The outage starts and stops exactly when the cp (or other access) starts and stops. This doesn't seem to be an MBUF problem, and I haven't been able to find a specific mention of this sort of problem. Is this something inherent with the kernel, the IDE driver, the fxp driver, or FreeBSD's SMP implementation? Or is this most likely an IRQ conflict or bad IRQ choice for the ether card? I plan on upgrading to 3.1R soon, so IRQ twiddling will be possible. If anyone has any insight I would _greatly_ appreciate it. I'll summarize if necessary. Thanks, -- Ken Brownfield kenb@irridia.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199904060731.CAA10029>