Date: Mon, 4 Jun 2007 16:43:45 -0400 From: "Vadym Chepkov" <vchepkov@gmail.com> To: "Max Laier" <max@love2party.net>, "Hugo Koji Kobayashi" <koji@registro.br> Cc: freebsd-pf@freebsd.org Subject: Re: udp fragmentation Message-ID: <004401c7a6e9$0c18ef60$050a0a0a@chepkov.lan> References: <20070528224225.GC40678@registro.br><200706021704.53787.max@love2party.net><20070604194430.GD21681@registro.br> <200706042200.14860.max@love2party.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Max,
This is exactly the same problem I have experienced before and I wrote it in
"Scrub problem" note on April, 14
I see amanda packets get lost after normalization and you are right, this is
exactly what happening - bad checksum for reassembled UDP packets:
$ netstat -ssp udp
udp:
14461468 datagrams received
38 with bad checksum
470 with no checksum
56858 dropped due to no socket
182267 broadcast/multicast datagrams dropped due to no socket
14222305 delivered
24009318 datagrams output
Jun 4 16:38:24 gateway kernel: pf_normalize_ip: reass frag 65286 @ 0-1480
Jun 4 16:38:24 gateway kernel: pf_normalize_ip: reass frag 65286 @
1480-2337
Jun 4 16:38:24 gateway kernel: pf_reassemble: 2337 < 2337?
Jun 4 16:38:24 gateway kernel: pf_reassemble: complete: 0xc4eb1300(2357)
$ netstat -ssp udp
udp:
14461572 datagrams received
39 with bad checksum
470 with no checksum
56858 dropped due to no socket
182273 broadcast/multicast datagrams dropped due to no socket
14222402 delivered
24009422 datagrams output
Sincerely,
Vadym Chepkov
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?004401c7a6e9$0c18ef60$050a0a0a>