On Jun 10, 2012, at 10:45, Jim Reid wrote: > On 10 Jun 2012, at 09:19, DTNX Postmaster wrote: > >> What type of queries? > > ANY queries for ihren.org with no UDP checksum: > > shaun# tcpdump -vv -n port 53 > 09:32:30.139803 IP (tos 0x0, ttl 251, id 24876, offset 0, flags [none], proto > UDP (17), length 66) 37.221.160.125.28832 > 93.186.33.42.53: [no cksum] > 18554+ [1au] ANY? ihren.org. ar: . OPT UDPsize=9000 (38) > 09:32:30.139806 IP (tos 0x0, ttl 251, id 24877, offset 0, flags [none], proto > UDP (17), length 66) 37.221.160.125.28832 > 93.186.33.42.53: [no cksum] > 18554+ [1au] ANY? ihren.org. ar: . OPT UDPsize=9000 (38) > 09:32:30.139929 IP (tos 0x0, ttl 251, id 24878, offset 0, flags [none], proto > UDP (17), length 66) 37.221.160.125.28832 > 93.186.33.42.53: [no cksum] > 18554+ [1au] ANY? ihren.org. ar: . OPT UDPsize=9000 (38)
This is what our tcpdump looks like; 11:20:46.115011 IP (tos 0x0, ttl 47, id 0, offset 0, flags [DF], proto UDP (17), length 63) 184.105.175.202.30632 > amonhen.nickserf.nl.domain: [udp sum ok] 43127+ ANY? xxxxxxxxxxxxx.tld. (35) 11:20:47.093295 IP (tos 0x0, ttl 47, id 0, offset 0, flags [DF], proto UDP (17), length 63) 184.105.175.202.50833 > amonhen.nickserf.nl.domain: [udp sum ok] 37318+ ANY? xxxxxxxxxxxxx.tld. (35) 11:20:48.290580 IP (tos 0x0, ttl 47, id 0, offset 0, flags [DF], proto UDP (17), length 63) 184.105.175.202.30559 > amonhen.nickserf.nl.domain: [udp sum ok] 24439+ ANY? xxxxxxxxxxxxx.tld. (35) 11:20:48.582575 IP (tos 0x0, ttl 47, id 0, offset 0, flags [DF], proto UDP (17), length 63) 184.105.175.202.53576 > amonhen.nickserf.nl.domain: [udp sum ok] 18641+ ANY? xxxxxxxxxxxxx.tld. (35) 11:20:48.993361 IP (tos 0x0, ttl 47, id 0, offset 0, flags [DF], proto UDP (17), length 63) 184.105.175.202.58969 > amonhen.nickserf.nl.domain: [udp sum ok] 23014+ ANY? xxxxxxxxxxxxx.tld. (35) (target domain obscured) >> The iptables rules mentioned in the first comment work well for us > > Well for starters, I [dw]on't use Linux. The server runs FreeBSD. Besides, > the damage is done by the time these packets hit the server's ethernet card. > At ~4000qps inbound, this is close to saturating the server's VLAN in the > data centre. The traffic needs to be blocked before it reaches that. I've > hopefully got the offending addresses blackholed by the name server now: > don't know though if those addresses were spoofed or not. I wasn't assuming you are using the same platform, just sharing our experiences of the past week :-) Perhaps it's possible to implement a similar rule in IPFW/PF/IPF? That way you at least won't be responding to the queries, and dropping the packets should lighten the load? Cya, Jona _______________________________________________ dns-operations mailing list dns-operations@lists.dns-oarc.net https://lists.dns-oarc.net/mailman/listinfo/dns-operations dns-jobs mailing list https://lists.dns-oarc.net/mailman/listinfo/dns-jobs