Discussion:
kern/144917: Flowtable crashes system
(too old to reply)
Evgenii Davidov
2010-03-21 09:04:55 UTC
Permalink
Здравствуйте,
It seems like flowtable has been merged and enabled by default in 8.0.... which is a really really bad idea.
On a system which handles two full BGP tables it makes one of the CPU cores run at 100% right after most of the prefixes get installed in the routing table.
i saw the same effect with ospf
--
Evgenii V Davidov
Ilya Zhuravlev
2010-03-25 15:32:45 UTC
Permalink
Post by Evgenii Davidov
Здравствуйте,
It seems like flowtable has been merged and enabled by default in 8.0.... which is a really really bad idea.
On a system which handles two full BGP tables it makes one of the CPU cores run at 100% right after most of the prefixes get installed in the routing table.
i saw the same effect with ospf
8.0-p2, 2 full-view with openbgpd
"tuning":
net.inet.tcp.blackhole=2
net.inet.udp.blackhole=1
net.inet.icmp.icmplim_output=0
net.inet.icmp.drop_redirect=1
net.inet.flowtable.nmbflows=32768

1 week uptime.Now I think only about increasing tx/rx descriptors to
reduce interrupts (default values was not changed)


netstat -w1 -Iigb0
input (igb0) output
packets errs bytes packets errs bytes colls
49100 0 12290513 23693 0 27268884 0
48322 0 12688283 24332 0 28099404 0
50602 0 12759620 24437 0 27698341 0
47857 0 11354124 21410 0 23845155 0

netstat -w1 -Iigb1
input (igb1) output
packets errs bytes packets errs bytes colls
32428 0 35027019 24562 0 5624934 0
30621 0 33384339 23569 0 4456944 0
28419 0 31014269 21571 0 3638083 0
29409 0 32524760 22137 0 3503600 0
30965 0 33532742 23973 0 5089231 0

netstat -w1 -Iem0
input (em0) output
packets errs bytes packets errs bytes colls
17217 0 3929366 72741 0 46377762 0
17412 0 3745112 75522 0 49338883 0
18385 0 4014568 77444 0 50532101 0
17142 0 3875518 77125 0 47646681 0
16870 0 3528316 73188 0 47940959 0
17069 0 3682891 80268 0 52904747 0
17313 0 4101576 75586 0 51933330 0
K. Macy
2010-04-03 03:07:35 UTC
Permalink
Please try with the latest 8-STABLE and tell me if recent changes fix it.

Thanks,
Kip
Post by Ilya Zhuravlev
Post by Evgenii Davidov
Здравствуйте,
It seems like flowtable has been merged and enabled by default in 8.0....
which is a really really bad idea.
On a system which handles two full BGP tables it makes one of the CPU
cores run at 100% right after most of the prefixes get installed in the
routing table.
i saw the same effect with ospf
8.0-p2, 2 full-view with openbgpd
net.inet.tcp.blackhole=2
net.inet.udp.blackhole=1
net.inet.icmp.icmplim_output=0
net.inet.icmp.drop_redirect=1
net.inet.flowtable.nmbflows=32768
1 week uptime.Now I think only about increasing tx/rx descriptors to reduce
interrupts (default values was not changed)
netstat -w1 -Iigb0
           input         (igb0)           output
  packets  errs      bytes    packets  errs      bytes colls
    49100     0   12290513      23693     0   27268884     0
    48322     0   12688283      24332     0   28099404     0
    50602     0   12759620      24437     0   27698341     0
    47857     0   11354124      21410     0   23845155     0
netstat -w1 -Iigb1
           input         (igb1)           output
  packets  errs      bytes    packets  errs      bytes colls
    32428     0   35027019      24562     0    5624934     0
    30621     0   33384339      23569     0    4456944     0
    28419     0   31014269      21571     0    3638083     0
    29409     0   32524760      22137     0    3503600     0
    30965     0   33532742      23973     0    5089231     0
netstat -w1 -Iem0
           input          (em0)           output
  packets  errs      bytes    packets  errs      bytes colls
    17217     0    3929366      72741     0   46377762     0
    17412     0    3745112      75522     0   49338883     0
    18385     0    4014568      77444     0   50532101     0
    17142     0    3875518      77125     0   47646681     0
    16870     0    3528316      73188     0   47940959     0
    17069     0    3682891      80268     0   52904747     0
    17313     0    4101576      75586     0   51933330     0
_______________________________________________
http://lists.freebsd.org/mailman/listinfo/freebsd-net
Barney Cordoba
2010-04-08 12:07:29 UTC
Permalink
Subject: Re: kern/144917: Flowtable crashes system
Date: Friday, April 2, 2010, 11:07 PM
Please try with the latest 8-STABLE
and tell me if recent changes fix it.
Thanks,
Kip
Post by Ilya Zhuravlev
Post by Evgenii Davidov
Здравствуйте,
On Sat, Mar 20, 2010 at 11:06:35PM +0000, Doychin
It seems like flowtable has been merged and
enabled by default in 8.0....
Post by Ilya Zhuravlev
Post by Evgenii Davidov
which is a really really bad idea.
On a system which handles two full BGP tables
it makes one of the CPU
Post by Ilya Zhuravlev
Post by Evgenii Davidov
cores run at 100% right after most of the
prefixes get installed in the
Post by Ilya Zhuravlev
Post by Evgenii Davidov
routing table.
i saw the same effect with ospf
8.0-p2, 2 full-view with openbgpd
net.inet.tcp.blackhole=2
net.inet.udp.blackhole=1
net.inet.icmp.icmplim_output=0
net.inet.icmp.drop_redirect=1
net.inet.flowtable.nmbflows=32768
1 week uptime.Now I think only about increasing tx/rx
descriptors to reduce
Post by Ilya Zhuravlev
interrupts (default values was not changed)
netstat -w1 -Iigb0
           input         (igb0)        
  output
Post by Ilya Zhuravlev
  packets  errs      bytes    packets  errs  
   bytes colls
Post by Ilya Zhuravlev
    49100     0   12290513      23693     0
  27268884     0
Post by Ilya Zhuravlev
    48322     0   12688283      24332     0
  28099404     0
Post by Ilya Zhuravlev
    50602     0   12759620      24437     0
  27698341     0
Post by Ilya Zhuravlev
    47857     0   11354124      21410     0
  23845155     0
Post by Ilya Zhuravlev
netstat -w1 -Iigb1
           input         (igb1)        
  output
Post by Ilya Zhuravlev
  packets  errs      bytes    packets  errs  
   bytes c
    32428     0   35027019      24562     0
   5624934     0
Post by Ilya Zhuravlev
    30621     0   33384339      23569     0
   4456944     0
Post by Ilya Zhuravlev
    28419     0   31014269      21571     0
   3638083     0
Post by Ilya Zhuravlev
    29409     0   32524760      22137     0
   3503600     0
Post by Ilya Zhuravlev
    30965     0   33532742      23973     0
   5089231     0
Post by Ilya Zhuravlev
netstat -w1 -Iem0
           input          (em0)        
  output
Post by Ilya Zhuravlev
  packets  errs      bytes    packets  errs  
   bytes colls
Post by Ilya Zhuravlev
    17217     0    3929366      72741     0
  46377762     0
Post by Ilya Zhuravlev
    17412     0    3745112      75522     0
  49338883     0
Post by Ilya Zhuravlev
    18385     0    4014568      77444     0
  50532101     0
Post by Ilya Zhuravlev
    17142     0    3875518      77125     0
  47646681     0
Post by Ilya Zhuravlev
    16870     0    3528316      73188     0
  47940959     0
Post by Ilya Zhuravlev
    17069     0    3682891      80268     0
  52904747     0
Post by Ilya Zhuravlev
    17313     0    4101576      75586     0
  51933330     0
Post by Ilya Zhuravlev
_______________________________________________
How about telling us how to turn it off; or better yet how to not
compile it into the kernel at all. Thats the best solution.

As my Dad used to say on a regular basis, we need this like we
need a hole in our head. Is 8.0 the Kip Macy personal test bed?

Why is something that virtually no-one needs enabled by default?

Barney
Vincent Hoffman
2010-04-08 13:05:34 UTC
Permalink
Post by Barney Cordoba
Subject: Re: kern/144917: Flowtable crashes system
Date: Friday, April 2, 2010, 11:07 PM
Please try with the latest 8-STABLE
and tell me if recent changes fix it.
Thanks,
Kip
Post by Ilya Zhuravlev
Post by Evgenii Davidov
Здравствуйте,
On Sat, Mar 20, 2010 at 11:06:35PM +0000, Doychin
It seems like flowtable has been merged and
enabled by default in 8.0....
Post by Ilya Zhuravlev
Post by Evgenii Davidov
which is a really really bad idea.
On a system which handles two full BGP tables
it makes one of the CPU
Post by Ilya Zhuravlev
Post by Evgenii Davidov
cores run at 100% right after most of the
prefixes get installed in the
Post by Ilya Zhuravlev
Post by Evgenii Davidov
routing table.
i saw the same effect with ospf
8.0-p2, 2 full-view with openbgpd
net.inet.tcp.blackhole=2
net.inet.udp.blackhole=1
net.inet.icmp.icmplim_output=0
net.inet.icmp.drop_redirect=1
net.inet.flowtable.nmbflows=32768
1 week uptime.Now I think only about increasing tx/rx
descriptors to reduce
Post by Ilya Zhuravlev
interrupts (default values was not changed)
netstat -w1 -Iigb0
input (igb0)
output
Post by Ilya Zhuravlev
packets errs bytes packets errs
bytes colls
Post by Ilya Zhuravlev
49100 0 12290513 23693 0
27268884 0
Post by Ilya Zhuravlev
48322 0 12688283 24332 0
28099404 0
Post by Ilya Zhuravlev
50602 0 12759620 24437 0
27698341 0
Post by Ilya Zhuravlev
47857 0 11354124 21410 0
23845155 0
Post by Ilya Zhuravlev
netstat -w1 -Iigb1
input (igb1)
output
Post by Ilya Zhuravlev
packets errs bytes packets errs
bytes colls
Post by Ilya Zhuravlev
32428 0 35027019 24562 0
5624934 0
Post by Ilya Zhuravlev
30621 0 33384339 23569 0
4456944 0
Post by Ilya Zhuravlev
28419 0 31014269 21571 0
3638083 0
Post by Ilya Zhuravlev
29409 0 32524760 22137 0
3503600 0
Post by Ilya Zhuravlev
30965 0 33532742 23973 0
5089231 0
Post by Ilya Zhuravlev
netstat -w1 -Iem0
input (em0)
output
Post by Ilya Zhuravlev
packets errs bytes packets errs
bytes colls
Post by Ilya Zhuravlev
17217 0 3929366 72741 0
46377762 0
Post by Ilya Zhuravlev
17412 0 3745112 75522 0
49338883 0
Post by Ilya Zhuravlev
18385 0 4014568 77444 0
50532101 0
Post by Ilya Zhuravlev
17142 0 3875518 77125 0
47646681 0
Post by Ilya Zhuravlev
16870 0 3528316 73188 0
47940959 0
Post by Ilya Zhuravlev
17069 0 3682891 80268 0
52904747 0
Post by Ilya Zhuravlev
17313 0 4101576 75586 0
51933330 0
Post by Ilya Zhuravlev
_______________________________________________
How about telling us how to turn it off; or better yet how to not
compile it into the kernel at all. Thats the best solution.
As my Dad used to say on a regular basis, we need this like we
need a hole in our head. Is 8.0 the Kip Macy personal test bed?
Why is something that virtually no-one needs enabled by default?
Barney
I havent followed the whole thread so if this doesnt actally work for
some reason then sorry for noise.

[***@prawn ~]# sysctl -a | grep flowtable
net.inet.ip.output_flowtable_size: 2048
net.inet.flowtable.nmbflows: 4096
net.inet.flowtable.tcp_expire: 86400
net.inet.flowtable.fin_wait_expire: 600
net.inet.flowtable.udp_expire: 300
net.inet.flowtable.syn_expire: 300
net.inet.flowtable.collisions: 0
net.inet.flowtable.max_depth: 0
net.inet.flowtable.free_checks: 8845
net.inet.flowtable.frees: 2181
net.inet.flowtable.misses: 2403
net.inet.flowtable.lookups: 11633
net.inet.flowtable.hits: 9231
net.inet.flowtable.enable: 1
net.inet.flowtable.debug: 0


[***@prawn ~]# sysctl -d net.inet.flowtable.enable
net.inet.flowtable.enable: enable flowtable caching.

[***@prawn ~]# sysctl net.inet.flowtable.enable=0
net.inet.flowtable.enable: 1 -> 0


Vince
Post by Barney Cordoba
_______________________________________________
http://lists.freebsd.org/mailman/listinfo/freebsd-net
K. Macy
2010-04-08 18:07:52 UTC
Permalink
Hi Vincent,
Thanks for your response. However, the answers to both his questions
have been given many times before and he knows them already. This
person has not made material contributions to discussions and has in
fact made active efforts to reduce the signal to noise ratio.

Cheers,
Kip
Post by Vincent Hoffman
Post by Barney Cordoba
Subject: Re: kern/144917: Flowtable crashes system
Date: Friday, April 2, 2010, 11:07 PM
Please try with the latest 8-STABLE
and tell me if recent changes fix it.
Thanks,
Kip
Post by Ilya Zhuravlev
Post by Evgenii Davidov
Здравствуйте,
On Sat, Mar 20, 2010 at 11:06:35PM +0000, Doychin
It seems like flowtable has been merged and
enabled by default in 8.0....
Post by Ilya Zhuravlev
Post by Evgenii Davidov
which is a really really bad idea.
On a system which handles two full BGP tables
it makes one of the CPU
Post by Ilya Zhuravlev
Post by Evgenii Davidov
cores run at 100% right after most of the
prefixes get installed in the
Post by Ilya Zhuravlev
Post by Evgenii Davidov
routing table.
i saw the same effect with ospf
8.0-p2, 2 full-view with openbgpd
net.inet.tcp.blackhole=2
net.inet.udp.blackhole=1
net.inet.icmp.icmplim_output=0
net.inet.icmp.drop_redirect=1
net.inet.flowtable.nmbflows=32768
1 week uptime.Now I think only about increasing tx/rx
descriptors to reduce
Post by Ilya Zhuravlev
interrupts (default values was not changed)
netstat -w1 -Iigb0
           input         (igb0)
  output
Post by Ilya Zhuravlev
  packets  errs      bytes    packets  errs
   bytes colls
Post by Ilya Zhuravlev
    49100     0   12290513      23693     0
  27268884     0
Post by Ilya Zhuravlev
    48322     0   12688283      24332     0
  28099404     0
Post by Ilya Zhuravlev
    50602     0   12759620      24437     0
  27698341     0
Post by Ilya Zhuravlev
    47857     0   11354124      21410     0
  23845155     0
Post by Ilya Zhuravlev
netstat -w1 -Iigb1
           input         (igb1)
  output
Post by Ilya Zhuravlev
  packets  errs      bytes    packets  errs
   bytes colls
Post by Ilya Zhuravlev
    32428     0   35027019      24562     0
   5624934     0
Post by Ilya Zhuravlev
    30621     0   33384339      23569     0
   4456944     0
Post by Ilya Zhuravlev
    28419     0   31014269      21571     0
   3638083     0
Post by Ilya Zhuravlev
    29409     0   32524760      22137     0
   3503600     0
Post by Ilya Zhuravlev
    30965     0   33532742      23973     0
   5089231     0
Post by Ilya Zhuravlev
netstat -w1 -Iem0
           input          (em0)
  output
Post by Ilya Zhuravlev
  packets  errs      bytes    packets  errs
   bytes colls
Post by Ilya Zhuravlev
    17217     0    3929366      72741     0
  46377762     0
Post by Ilya Zhuravlev
    17412     0    3745112      75522     0
  49338883     0
Post by Ilya Zhuravlev
    18385     0    4014568      77444     0
  50532101     0
Post by Ilya Zhuravlev
    17142     0    3875518      77125     0
  47646681     0
Post by Ilya Zhuravlev
    16870     0    3528316      73188     0
  47940959     0
Post by Ilya Zhuravlev
    17069     0    3682891      80268     0
  52904747     0
Post by Ilya Zhuravlev
    17313     0    4101576      75586     0
  51933330     0
Post by Ilya Zhuravlev
_______________________________________________
How about telling us how to turn it off; or better yet how to not
compile it into the kernel at all. Thats the best solution.
As my Dad used to say on a regular basis, we need this like we
need a hole in our head. Is 8.0 the Kip Macy personal test bed?
Why is something that virtually no-one needs enabled by default?
Barney
I havent followed the whole thread so if this doesnt actally work for
some reason then sorry for noise.
net.inet.ip.output_flowtable_size: 2048
net.inet.flowtable.nmbflows: 4096
net.inet.flowtable.tcp_expire: 86400
net.inet.flowtable.fin_wait_expire: 600
net.inet.flowtable.udp_expire: 300
net.inet.flowtable.syn_expire: 300
net.inet.flowtable.collisions: 0
net.inet.flowtable.max_depth: 0
net.inet.flowtable.free_checks: 8845
net.inet.flowtable.frees: 2181
net.inet.flowtable.misses: 2403
net.inet.flowtable.lookups: 11633
net.inet.flowtable.hits: 9231
net.inet.flowtable.enable: 1
net.inet.flowtable.debug: 0
net.inet.flowtable.enable: enable flowtable caching.
net.inet.flowtable.enable: 1 -> 0
Vince
Post by Barney Cordoba
_______________________________________________
http://lists.freebsd.org/mailman/listinfo/freebsd-net
_______________________________________________
http://lists.freebsd.org/mailman/listinfo/freebsd-net
Loading...