Date: Mon, 1 Mar 1999 07:21:52 -0800 (PST) From: Matthew Jacob <mjacob@feral.com> To: Matthew Dillon <dillon@apollo.backplane.com> Cc: freebsd-hackers@FreeBSD.ORG Subject: Re: Panic in FFS/4.0 as of yesterday - update Message-ID: <Pine.LNX.4.04.9903010715310.14256-100000@feral-gw> In-Reply-To: <199903010810.AAA42400@apollo.backplane.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Most Very Excellent- I'll give it all a try. One comment: > > Old Bugs not fixed: > > * I/O saturation is still a problem. There is no easy solution - > even reverting to synchronous I/O doesn't help because the STEST > script starts up 50 processes doing writes. This isn't a serious > bug under normal operating conditions but it is annoying. > This is a less likely real-world scenario- you don't usually start up multiple things all at once which then suck down the entire buffer cache. Would the problem manifest itself under increasing load? One thing I'm mulling doing is to try and move forward musbus or it's equivalent- eons ago during the initial Solaris development we used musbus and it's ilk to generate stepwise increasing load to find where the kernel broke. Similarily, it may not be the instantaneous load of 50 or 100 writers springing into action that drags a system to unusable, but the slowly increasing load. I'll ponder a couple of ways to make this all a test package at the very least- I don't know about you, but I find this stuff helpful. Oh- I take it from your found but not fixed mail that doreallocblks should set to zero still.... Thanks again, Matt- -matt (the lower case matt). To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.4.04.9903010715310.14256-100000>