Date: Tue, 4 Jun 2019 01:27:11 +0700 From: Eugene Grosbein <eugen@grosbein.net> To: Larry Rosenman <ler@lerctr.org> Cc: Freebsd net <freebsd-net@freebsd.org>, owner-freebsd-net@freebsd.org Subject: Re: SNMP/No Bufferspace Message-ID: <2ada895d-fbeb-077e-e116-5d9837ce0cb7@grosbein.net> In-Reply-To: <03051c226dadc9907dcd0839d5025e18@lerctr.org> References: <fde5e5f1b69c85fd5c7a7ab12e3d7167@lerctr.org> <f2cdca32-2df1-7ad0-b5c1-cf3bc4905fa9@grosbein.net> <8dba5eaea800356d279eb81d7d8b2a30@lerctr.org> <feed38df-c85c-6643-8382-c12366576765@grosbein.net> <03051c226dadc9907dcd0839d5025e18@lerctr.org>
next in thread | previous in thread | raw e-mail | index | archive | help
04.06.2019 1:15, Larry Rosenman wrote: > Name Mtu Network Address Ipkts Ierrs Idrop Opkts Oerrs Coll Drop > bce0 1.5K <Link#2> a4:ba:db:29:66:95 715M 186k 0 533M 0 0 15M Something really bad happening at driver level or physical level: way too many input errors (perhaps, bad CRC) and output drops (due to busy media?). I prefer to max out interface FIFO length while using bce(4) by means of /boot/loader.conf when link is heavily utilized: # max = 8*255 = 2040 hw.bce.tx_pages=8 Also, check driver-level stats with sysctl dev.bce.0 If possible, check link partner - managed switch port counters or what does it have. Maybe, replace twisted pair cable.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?2ada895d-fbeb-077e-e116-5d9837ce0cb7>