Date: Wed, 16 Jan 2002 10:49:04 +0000 From: Rasputin <rasputin@submonkey.net> To: Matthew Dillon <dillon@apollo.backplane.com> Cc: stable@freebsd.org Subject: Re: what's vnlru? Message-ID: <20020116104904.A2800@shikima.mine.nu> In-Reply-To: <200201160101.g0G117r64693@apollo.backplane.com>; from dillon@apollo.backplane.com on Tue, Jan 15, 2002 at 05:01:07PM -0800 References: <200201151618468.SM01176@141.com> <200201160101.g0G117r64693@apollo.backplane.com>
next in thread | previous in thread | raw e-mail | index | archive | help
* Matthew Dillon <dillon@apollo.backplane.com> [020116 01:10]: > vnlru is responsible for flushing and freeing vnodes when you > hit the kern.maxvnodes limit. The history of this is: > > 4.4 and earlier vnodes were only freed if they had no cached pages. > This could lead to a vnode blowout on machines with > large amounts of ram accessing lots of small files. > (the kern.maxvnodes limit was essentially ignored) > > 4.4-stable I introduced code to flush/reuse vnodes on the fly, > but this lead to possible locking deadlocks. > > 4.5-* Peter introduced a separate kernel thread to deal with > vnode flushing issues, 'vnlru', in order to fix the > possible locking deadlocks. The on-the-fly code > was changed to only be called from the kernel thread. > > This kernel thread will generally not be very active, even on a heavily > loaded system. You have to have a lot of ram and be accessing tens of > thousands of tiny files to activate the thread. Thansk for the overview - it's not a performance issue, I just checked another fairly recent (4.5 pre) BSD box and didn't see it, so thought I'd better check it out. [ Is this the sort of thing that should go into UPDATING? ] Thanks to all ! -- Learned men are the cisterns of knowledge, not the fountainheads. Rasputin :: Jack of All Trades - Master of Nuns :: To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020116104904.A2800>