sendto: No buffer space

anujshrestha

New Member


Messages: 1

hiii,
i m getting these in /var/log/messeges due to which it looks like causing the service time trouble in squid, Do anybody experienced this.




Code:
Jan 16 12:58:13 proxy2 syslogd: sendto: No buffer space available
Jan 16 12:58:13 proxy2 last message repeated 13 times
Jan 16 13:01:32 proxy2 syslogd: sendto: No buffer space available
Jan 16 13:01:32 proxy2 last message repeated 33 times
Jan 16 13:02:44 proxy2 last message repeated 48 times
Jan 16 13:05:32 proxy2 last message repeated 6 times
Code:
proxy2# uname -a
FreeBSD proxy2.wlink.com.np 7.0-RELEASE FreeBSD 7.0-RELEASE #0: Wed Nov 12 18:26:19 NPT 2008     [email]root@proxy2.wlink.com.np[/email]:
/usr/obj/usr/src/sys/mykern  i386
 

kamikaze

Well-Known Member

Reaction score: 75
Messages: 365

Are you using a slow connection? It appears your TCP-buffer has filled up to the max with requests and cannot collect any more messages. You can increase the maximum buffer size:

net.inet.tcp.recvbuf_max
net.inet.tcp.sendbuf_max

But this can really just help you over short traffic spikes. I think your connection is simply too slow to handle the traffic.
 

Const

Member


Messages: 35

I have the same problem. Server runs for a few minutes and stop responding after that. Trying to ping anyone from the server give me
Code:
# ping 192.168.1.1
PING 192.168.1.1 (192.168.1.1): 56 data bytes
ping: sendto: No buffer space available
ping: sendto: No buffer space available
My kernel recompiled with maxusers set to 512 and nmbclusters to 65532, also options IPFIREWALL_FORWARD and IPFIREWALL_NAT are included. The OS version is
Code:
# uname -a
FreeBSD  7.2-RELEASE FreeBSD 7.2-RELEASE #1: Fri May 15 15:46:53 MSD 2009     const@:/usr/src/sys/amd64/compile/MYKERNEL-IPFW-NAT  amd64
Here is my ipfw rules (80 port forwards to squid 2.7, the rest traffic goes through nat)
Code:
# ipfw show
00100     0      0 check-state
00200     0      0 allow ip from any to any via lo0
00300     0      0 allow log logamount 10000 ip from 192.168.1.10 to any
00400     0      0 allow log logamount 10000 ip from any to 192.168.1.10
00500     0      0 deny ip from 192.168.2.39 to any dst-port 25
00600     1     40 fwd 192.168.1.2,3128 log logamount 10000 ip from 192.168.2.0/24 to any dst-port 80 in via 192.168.2.6
00700     0      0 allow log logamount 10000 ip from 192.168.1.2 to any dst-port 80 out via 192.168.1.2
00800     0      0 allow log logamount 10000 ip from any 80 to 192.168.1.2 in via 192.168.1.2
00900    17  19218 allow log logamount 10000 ip from any 80 to 192.168.2.0/24 out via 192.168.2.6
01000     0      0 allow log logamount 10000 ip from 192.168.2.0/24 to me dst-port 22 setup keep-state
01100  2571 380532 nat 1 log logamount 10000 ip from any to any established
01200     0      0 allow log logamount 10000 udp from 192.168.1.2 to 192.168.1.1 dst-port 53
01300     0      0 allow log logamount 10000 udp from 192.168.1.1 53 to 192.168.1.2 in via 192.168.1.2
01400     0      0 allow icmp from 192.168.2.0/24 to me in via 192.168.2.6
01500     0      0 allow icmp from me to 192.168.2.0/24 out via 192.168.2.6
01600     0      0 nat 1 log logamount 10000 ip from 192.168.2.0/24 to any dst-port 443 setup keep-state
01700     3    144 nat 1 log logamount 10000 ip from 192.168.2.0/24 to any dst-port 25,110 setup keep-state
01800     0      0 nat 1 log logamount 10000 ip from 192.168.2.0/24 to any dst-port 5190 setup keep-state
01900     0      0 nat 1 log logamount 10000 ip from 192.168.2.0/24{40-42,100} to 86.111.5.21 dst-port 9118 setup keep-state
02000     0      0 nat 1 log logamount 10000 ip from 192.168.2.85 to 195.68.252.37 dst-port 1352 setup keep-state
02100     0      0 nat 1 log logamount 10000 ip from 192.168.2.0/24{21,24,30-33,240,241} to any setup keep-state
65000     2     96 deny log logamount 10000 tcp from any to any
65100 10199 896488 deny ip from any to any
65535    10    500 deny ip from any to any
192.168.2.0/24 - my local net
192.168.1.1 - router (ISP is the next hop after it)
192.168.2.6 - internal interface
192.168.1.2 - external

I have about 100 users in local net using this proxy. I doubt the problem goes to slow connection speed, users just serfing. But even if it's possible shouldn't network card or OS or whatever responsible for packets transfering just drop packets if it can't handle them and not to hang all the network activity?
 

Const

Member


Messages: 35

I'm an idiot... Looks like network card just slightly fell out from slot and connection's become unstable. I'll test it now and post if the problem isn't solved
 
Top