|
Actually I'm not sure what you mean by the ping test. Things haven't gotten
Well, you did a broadcast ping once that seemed to have triggered the
table overflow.
bad enough to get the buffer space error. I still see the Neighbour table
overflow error in dmesg, but these were already there from before changing
gc_thresh3
Any new ones?
Here are all the diags you wanted me to run earlier.
Can you add some more please?
ip -o -s route show cache | wc -l
grep cache /proc/slabinfo
The latter instead of cat /proc/slabinfo
-------------------------------------------------
ip addr show
1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
2: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 100
link/ether 00:01:fa:ff:fc:45 brd ff:ff:ff:ff:ff:ff
inet 216.163.120.8/24 brd 216.163.120.255 scope global eth0
inet 216.163.120.45/24 brd 216.163.120.255 scope global secondary eth0:0
inet 216.163.120.46/24 brd 216.163.120.255 scope global secondary eth0:1
inet 216.163.120.62/24 brd 216.163.120.255 scope global secondary eth0:2
inet 216.163.120.47/24 brd 216.163.120.255 scope global secondary eth0:3
inet 216.163.120.63/24 brd 216.163.120.255 scope global secondary eth0:4
inet 216.163.120.48/24 brd 216.163.120.255 scope global secondary eth0:5
inet 216.163.120.119/24 brd 216.163.120.255 scope global secondary
Ok, I haven't payed attention to this setup before but all those eth0:*
are supposed to be VIPs, right? If so they need to /32 and not /24. If
you have them /24 all of them will reply to any address. Please set a
VIP up like follows:
ip addr add 216.163.120.45/32 brd + dev eth0 label eth0:0
ip addr add 216.163.120.46/32 brd + dev eth0 label eth0:1
ip addr add 216.163.120.47/32 brd + dev eth0 label eth0:3
[...]
cat /proc/net/rt_cache_stat
00004846 016f041d 00255164 00000000 00000000 00000456 00000000 00000025
003daa6a 000ad632 0000045a 000d7622 000d68e9 0
000018e 00000000
00004846 016ece54 00254aaf 00000000 00000000 0000042c 00000000 00000020
003db1e3 000abcf9 000004ce 000d74e3 000d6716 0
0000167 00000000
-------------------------------------------------
ip_dst_cache 18760 20020 192 1001 1001 1 : 252 126
arp_cache 1560 1560 128 52 52 1 : 252 126
Pretty damn hard. Could you please run following script?
---------------------------------------------
#!/bin/sh
echo "30" >/proc/sys/net/ipv4/neigh/default/gc_interval
echo "20" >/proc/sys/net/ipv4/neigh/default/gc_stale_time
echo "512" >proc/sys/net/ipv4/neigh/default/gc_thresh1
echo "2048" >/proc/sys/net/ipv4/neigh/default/gc_thresh2
echo "4096" >/proc/sys/net/ipv4/neigh/default/gc_thresh3
echo "20" >/proc/sys/net/ipv4/neigh/eth0/gc_stale_time
echo "20" >/proc/sys/net/ipv4/neigh/eth1/gc_stale_time
echo "20" >/proc/sys/net/ipv4/neigh/lo/gc_stale_time
ip route flush cache
---------------------------------------------
Best regards,
Roberto Nibali, ratz
--
echo '[q]sa[ln0=aln256%Pln256/snlbx]sb3135071790101768542287578439snlbxq'|dc
| 