Re: error: dst cache overflow in kernel 2.3.15

Markus Hennig (hennig@planNET.de)
Tue, 31 Aug 1999 13:10:30 +0200

kuznet@ms2.inr.ac.ru wrote:
>
> Hello!
>
> > I'll try to reproduce this here and see happens...
>
> Please... My experience tells me that route leaks are almost
> impossible to find before I am able to reproduce them.
>
> Please, monitor ip_dst_cache in /proc/slabinfo, periodically
> flushing cache and try to guess the point, when it starts to grow.
> Yes, you can also make "ip -o route ls cache | wc" and subtract
> it from ip_dst_cache to get number of dead routes.
>
> Alexey
>
here is the log of the attemp to reproduce the error:

45 min
./ip -o route ls cache|wc -l 42
ip_dst_cache 55 75

12:04pm up 1:04, 10 users, load average: 0.08, 0.53, 0.80
56
ip_dst_cache 70 75

katherin:/tmp/990709/iproute2/ip # uptime; ./ip -o route ls cache|wc -l
;
cat /proc/slabinfo | grep ip_dst
12:19pm up 1:20, 11 users, load average: 0.08, 0.07, 0.30
35
ip_dst_cache 577 600

after restart of my fvwm2:

12:28pm up 1:29, 11 users, load average: 0.13, 0.15, 0.22
64
ip_dst_cache 607 625

12:51pm up 1:51, 12 users, load average: 0.92, 0.89, 0.50
43
ip_dst_cache 779 800

1:12pm up 2:13, 13 users, load average: 0.14, 0.05, 0.15
56
ip_dst_cache 891 900

1:44pm up 2:44, 13 users, load average: 0.09, 0.13, 0.11
54
ip_dst_cache 898 925

1:56pm up 2:56, 12 users, load average: 0.00, 0.04, 0.06
49
ip_dst_cache 1004 1025

2:17pm up 3:17, 12 users, load average: 0.00, 0.02, 0.05
57
ip_dst_cache 1062 1075

2:41pm up 3:41, 13 users, load average: 0.00, 0.02, 0.02
41
ip_dst_cache 1265 1275

2:51pm up 3:51, 13 users, load average: 0.09, 0.03, 0.01
56
ip_dst_cache 1280 1300

3:18pm up 4:18, 14 users, load average: 0.08, 0.17, 0.11
66
ip_dst_cache 1290 1300

3:21pm up 4:22, 15 users, load average: 0.01, 0.10, 0.09
63
ip_dst_cache 1287 1300

4:08pm up 5:08, 15 users, load average: 0.13, 0.04, 0.01
63
ip_dst_cache 1287 1300

after compile netfilter-0.1.5 (?????)

4:23pm up 5:23, 16 users, load average: 0.03, 0.13, 0.07
56
ip_dst_cache 1530 1550

5:04pm up 6:05, 21 users, load average: 0.21, 0.13, 0.09
55
ip_dst_cache 2073 2100

5:30pm up 6:30, 20 users, load average: 0.00, 0.03, 0.00
63
ip_dst_cache 2614 2625

5:41pm up 6:42, 19 users, load average: 0.34, 0.22, 0.09
57
ip_dst_cache 2700 2700

5:42pm up 6:42, 19 users, load average: 0.56, 0.27, 0.11
61
ip_dst_cache 2703 2725

5:51pm up 6:52, 19 users, load average: 0.06, 0.09, 0.09

58
ip_dst_cache 2872 2900

9:59am up 23:00, 19 users, load average: 0.01, 0.04, 0.01
55
ip_dst_cache 2912 2925

11:00am up 1 day, 0 min, 20 users, load average: 0.25, 0.10, 0.02
66
ip_dst_cache 3930 3950

11:30am up 1 day, 31 min, 20 users, load average: 0.10, 0.09, 0.07
116
ip_dst_cache 4035 4100

12:26pm up 1 day, 1:27, 20 users, load average: 0.81, 0.29, 0.11
116
ip_dst_cache 4096 4100

Aug 31 12:24:09 katherin kernel: dst cache overflow
Aug 31 12:24:14 katherin last message repeated 9 times
Aug 31 12:24:14 katherin kernel: NET: 8 messages suppressed.
Aug 31 12:24:14 katherin kernel: dst cache overflow
Aug 31 12:25:18 katherin kernel: NET: 10 messages suppressed.

echo 1 > /proc/sys/net/ipv4/route/flush
ip_dst_cache 4096 4100

nothing works

echo 1 > /proc/sys/net/ipv4/route/flush
ip_dst_cache 4096 4100

networking is agian working

12:30pm up 1 day, 1:30, 20 users, load average: 0.02, 0.14, 0.08
7
ip_dst_cache 4096 4100

12:31pm up 1 day, 1:32, 20 users, load average: 0.03, 0.12, 0.08
9
ip_dst_cache 4096 4100

all consisting connections are working, but new connection cant
established

i hope it help - for further test i wait for your instructions ;-)

MfG, Markus Hennig

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/