| 1 |
Kashani wrote: |
| 2 |
|
| 3 |
> I've been Googling for a few hours looking for information from |
| 4 |
>anyone actually using syn cookies in a production system. We'd been hit |
| 5 |
>with a few syn floods recently and were looking at a number of ways to |
| 6 |
>stop them. |
| 7 |
> |
| 8 |
> From my research it appears that syn cookies should not affect the |
| 9 |
>TCP session. There are some references to syn cookies causing slow downs |
| 10 |
>with SMTP and ftp, but nothing really concrete. Also most of the info is |
| 11 |
>from 2001 when 2.4 came out and syn cookies were linked with the newer |
| 12 |
>iptables code. Was looking for case studies and hard stats. |
| 13 |
> |
| 14 |
> I'd like to hear from anyone running web server farms that has syn |
| 15 |
>cookies turned on. Ideally you'd be pushing a fair amount of traffic or |
| 16 |
>have 1k concurrent users per server at some point so I know it'll scale. |
| 17 |
>Did you have any problems, weird firewalls blocking your servers, latency |
| 18 |
>connection to the server, load balancers didn't like it, weird connections |
| 19 |
>that hung around forever, and so forth. |
| 20 |
> |
| 21 |
>Thanks in advance, |
| 22 |
>kashani |
| 23 |
> |
| 24 |
> |
| 25 |
hi, |
| 26 |
|
| 27 |
We have syn cookies turned on since we had a attack (6 mbit) and without |
| 28 |
syn cookies off the load of the servers where very high login in was |
| 29 |
almost not possible. |
| 30 |
and when syn cookies turned on the load go down and the servers working |
| 31 |
ok slow but website where up again. |
| 32 |
and we have have no problems everything is normal. |
| 33 |
|
| 34 |
grzt rob ter haar |
| 35 |
NedLinux.nl |
Archives