Date: Fri, 7 Mar 2008 10:19:58 -0300 From: JoaoBR <joao@matik.com.br> To: freebsd-current@freebsd.org Cc: matthew <matthew@matthew.sk> Subject: Re: FreeBSD 7.0 Beta, RC, RELEASE (amd64) freezes with dummynet enabled Message-ID: <200803071019.58870.joao@matik.com.br> In-Reply-To: <47D12843.9080303@matthew.sk> References: <506309.93775.qm@web63903.mail.re1.yahoo.com> <47C879F1.6000407@matthew.sk> <47D12843.9080303@matthew.sk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Friday 07 March 2008 08:34:27 matthew wrote: > > I am bringing back to live this thread, please help me to resolve this > bug, some clue, where to look, i think this is a critical bug which > prevents to use 7.X on routers used for shaping. > > I have disabled polling but still, the box freezes after some minutes > under heavy traffic ~150-200Mbit/s. > I have tested the for network performance pool with iperf with load > ~150Mbit/s with 10 paralels streams, but the box didn`t freeze. > > The box is an fileserver, so maybe acces to disk operations may be a clue. > > I will also try it with ULE insteat of ULE, but i think in the past with > BETA2 i had the ULE schneduler. > I t seems I found a pattern on my machine the problem happens only with dynamic pipes using the=20 =2E.. mask src-ip|dst-ip 0xffffffff using individual pipes and bw definitions for each appearently the problem= =20 does not appear (here) I have not your traffic but here it freezes after a couple of hours or so this morning I needed to open the firewall and did it issuing ipfw add 1 pa= ss=20 ip from any to any and the server freezed imediately I could repeat it, opening the fw after some pipes are counting and the=20 machine freeze the same moment so far I didn't get the freeze with individual pipes I got one panic which pointed to pmap (curr proc cc anothertime syncer, I w= as=20 compiling the kernel) what is strange also, because almost ever compiling=20 something as kernel or world triggers the freeze but only when dynamic pipe= s=20 are underway and more strange only with SATA disks not on SCSI, without=20 dynamic pipes I can stress the disk as much as I want and it runs stable If i restart the firewall by flushing all rules first I can compile without= =20 problems I am sure this is no hardware thing because I changed it completly several= =20 times I am with latest sources and kernel/world from this night =2D-=20 Jo=E3o A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura. Service fornecido pelo Datacenter Matik https://datacenter.matik.com.br
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200803071019.58870.joao>