ping: Could only allocate a receive buffer of 8191 bytes (default 65535)

[Joao Ronaldo]

Hi,

I am running obsd 4.7.

I noticed the message in the subject whenever I use ping.

There is no packet loss but I am worried I might be reaching
some sort of resource exhaustion.

Could some body help me to interpret this message and
possibly give hints on how to avoid it?

Thanks in advance.

Regards,

JoC#o

[Christiano F. Haesbaert]

Post by Joao Ronaldo
Hi,
I am running obsd 4.7.

That's a bit old.

Post by Joao Ronaldo
I noticed the message in the subject whenever I use ping.
There is no packet loss but I am worried I might be reaching
some sort of resource exhaustion.
Could some body help me to interpret this message and
possibly give hints on how to avoid it?

setsockopt is failing in:

while (setsockopt(s, SOL_SOCKET, SO_RCVBUF,
(void*)&bufspace, sizeof(bufspace)) < 0) {
if ((bufspace -= 1024) <= 0)
err(1, "Cannot set the receive buffer size");
}
if (bufspace < IP_MAXPACKET)
warnx("Could only allocate a receive buffer of %i bytes (default %i)",
bufspace, IP_MAXPACKET);

Did you tweak any sysctl options ?

Is the machine loaded ?

I think sbchecklowmem() may be failing in uipc_socket2.c:403.

Also ping should use warn() instead of a warnx() there.

Claudio ?

[Joao Ronaldo]

Hi,

Thanks for your reply

Post by Christiano F. Haesbaert
Did you tweak any sysctl options ?

net.inet.ip.ifq.maxlen=512
net.inet.tcp.recvspace=131072
net.inet.tcp.sendspace=65536

Post by Christiano F. Haesbaert
B Is the machine loaded ?

Not at all

load averages: B 0.10, B 0.11, B 0.08 B B 11:13:56
51 processes: B 1 running, 49 idle, 1 on processor
CPU states: B 0.8% user, B 0.0% nice, B 0.2% system, B 0.2% interrupt, 98.9% idle
Memory: Real: 261M/400M act/tot B Free: 599M B Swap: 0K/2055M used/tot

B PID USERNAME PRI NICE B SIZE B RES STATE B B WAIT B B B TIME B B CPU COMMAND
20819 sshd B B B 2 B B 0 1988K 1972K sleep B B select B B 0:00 B 0.15% sshd
6710 root B B B 2 B B 0 1980K 2432K sleep B B netio B B 0:00 B 0.15% sshd
17426 _spamd B B 2 B B 0 B 17M B 11M sleep B B select B B 6:38 B 0.00% spamd
27968 ebl B B B B 2 B B 0 3872K 4544K sleep B B kqread B B 0:05 B 0.00% tmux

# netstat -m B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B 11:13:56
6486 mbufs in use:
B B B B 6417 mbufs allocated to data
B B B B 61 mbufs allocated to packet headers
B B B B 8 mbufs allocated to socket names and addresses
3250/3430/6144 mbuf 2048 byte clusters in use (current/peak/max)
0/8/6144 mbuf 4096 byte clusters in use (current/peak/max)
0/8/6144 mbuf 8192 byte clusters in use (current/peak/max)
34/42/6144 mbuf 9216 byte clusters in use (current/peak/max)
0/8/6144 mbuf 12288 byte clusters in use (current/peak/max)
0/8/6144 mbuf 16384 byte clusters in use (current/peak/max)
0/8/6144 mbuf 65536 byte clusters in use (current/peak/max)
9020 Kbytes allocated to network (93% in use)
0 requests for memory denied
0 requests for memory delayed
0 calls to protocol drain routines

Regards,

JoC#o.

[Christiano F. Haesbaert]

He fixed it by increasing kern.maxclusters.

Post by Joao Ronaldo
Hi,
I am running obsd 4.7.
I noticed the message in the subject whenever I use ping.
There is no packet loss but I am worried I might be reaching
some sort of resource exhaustion.
Could some body help me to interpret this message and
possibly give hints on how to avoid it?
Thanks in advance.
Regards,
JoC#o

[Robert]

Post by Christiano F. Haesbaert
He fixed it by increasing kern.maxclusters.

Thanks for including the solution, it helped me out!