Date: Sat, 8 Mar 2003 23:25:24 -0400 (AST) From: "Marc G. Fournier" <scrappy@hub.org> To: Tor.Egge@cvsup.no.freebsd.org Cc: freebsd-stable@freebsd.org Subject: Re: vlrureclaim update Message-ID: <20030308231951.X6638@hub.org> In-Reply-To: <20030308195134A.tegge@cvsup.no.freebsd.org> References: <20030308051800.G66674@hub.org> <20030308174057Y.tegge@cvsup.no.freebsd.org> <20030308153059.J6638@hub.org> <20030308195134A.tegge@cvsup.no.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 8 Mar 2003 Tor.Egge@cvsup.no.freebsd.org wrote: > I believe the vnlru_proc was intended as a fallback method for freeing > vnodes. A vnode is normally added to the free list when the last page > owned by a object of type OBJT_VNODE is freed, see vm_page_free_toq() in > vm_page.c. Was just thinking of this ... I just kill'd off 18 jails (and unmounted the associated union and proc fs's) on one of the servers, and right now, I'm seeing: debug.numvnodes: 234726 - debug.freevnodes: 3518 - debug.vnlru_nowhere: 2 - vlruwt debug.numvnodes: 234726 - debug.freevnodes: 3240 - debug.vnlru_nowhere: 2 - vlruwt debug.numvnodes: 234726 - debug.freevnodes: 3064 - debug.vnlru_nowhere: 2 - vlruwt Now, if the free vnodes are *supposed* to be auto-reclaiming, and teh vnlru_proc is meant to provide a fallback, then is there a bug with how vm_page_free_toq is working (or not being called, maybe?) ... ? Similar to the problem I had mentioend in relation to softupdates? Where the space for files being deleted didn't look like it was being reclaimed until the server is/was rebooted ... 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?20030308231951.X6638>