This is a discussion on How Do I block DNS Update Attempts within the Linux Security forums, part of the System Security and Security Related category; I keep having somebody attempting to Update our DNS Records. We do not know who this is and an NSLOOKUP ...
|
|||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
|
|||
|
I keep having somebody attempting to Update our DNS Records. We do not know
who this is and an NSLOOKUP tells us they are from a DHCP Pool owned by verizon.net. Calling Verszon is useless. Periodically our Linux Box needs to be rebooted due to lockups probably caused by these failed update attempts. Question: How do I block these update attempts. Non-authoritative answer: 29.236.156.141.in-addr.arpa name = pool-141-156-236-29.res.east.verizon.net. -Thank You -tom May 14 14:41:03 smtp2 named[967]: client 141.156.236.29#3114: updating zone 'mydomain.com/IN': update failed: 'RRset exists (value dependent)' prerequisite not satisfied (NXRRSET) May 14 14:41:03 smtp2 named[967]: client 141.156.236.29#3117: update 'mydomain.com/IN' denied May 14 14:41:47 smtp2 named[966]: client 141.156.236.29#3120: updating zone 'mydomain.com/IN': update failed: 'RRset exists (value dependent)' prerequisite not satisfied (NXRRSET) May 14 14:41:47 smtp2 named[967]: client 141.156.236.29#3123: update 'mydomain.com/IN' denied May 14 14:43:05 smtp2 sshd(pam_unix)[2074]: session closed for user root May 14 14:46:03 smtp2 named[966]: client 141.156.236.29#3126: updating zone 'mydomain.com/IN': update failed: 'RRset exists (value dependent)' prerequisite not satisfied (NXRRSET) May 14 14:46:03 smtp2 named[966]: client 141.156.236.29#3129: update 'mydomain.com/IN' denied May 14 14:46:47 smtp2 named[966]: client 141.156.236.29#3132: updating zone 'mydomain.com/IN': update failed: 'RRset exists (value dependent)' prerequisite not satisfied (NXRRSET) May 14 14:46:47 smtp2 named[966]: client 141.156.236.29#3135: update 'mydomain.com/IN' denied May 14 14:48:26 smtp2 sshd(pam_unix)[2153]: session opened for user root by (uid=0) [root@smtp2 root]# |
|
|||
|
"Tom Celica" <tom@dontreply.com> writes:
> I keep having somebody attempting to Update our DNS Records. We do not > know who this is and an NSLOOKUP tells us they are from a DHCP Pool owned > by verizon.net. Calling Verszon is useless. What relationship does your organization have with verizon? Any Windoze-lusing employees dialling-up from home accessing work, for example? Why does this worry you, anyway? As long as you've configured your nameserver properly (you don't say what named you're using, so RTFM for whatever it is), it's not as they'll be able to do anything, anyway, right? > Periodically our Linux Box needs to be rebooted due to lockups probably > caused by these failed update attempts. Highly unlikely. There are proper ways to diagnose machines hanging starting with correlating console output, logfiles (two events at 03 and 47s past the minute, 5mins apart - have you noticed this pattern?), system load, CPU and ambient temperature, whether motherboard components are well seated or not, quality of PSU and smoothness of input voltage etc, without resorting to voodoo guesswork. [snip] ~Tim -- There's a sadness, there's a joy |piglet@stirfried.vegetable.org.uk There's a place, |http://spodzone.org.uk/cesspit/ There's a song that will never die | Forever | |
|
|||
|
Tom Celica wrote:
> I keep having somebody attempting to Update our DNS Records. We do not know > who this is and an NSLOOKUP tells us they are from a DHCP Pool owned by > verizon.net. Calling Verszon is useless. > > Periodically our Linux Box needs to be rebooted due to lockups probably > caused by these failed update attempts. > > Question: How do I block these update attempts. > > Non-authoritative answer: > 29.236.156.141.in-addr.arpa name = > pool-141-156-236-29.res.east.verizon.net. > > You block them by configuring bind to ignore them. If you have bind configured properly, then no harm done. You could IPTABLE DNS requests from that IP range. Since it's dial-up, it's a fair bet that they use Verizon's RADIUS-supplied DNS settings and those DNS machines are probably not even close in address space (or one would hope). The concept being that since they are dial-up, the supplied DNS will do the look-up and isn't blocked. However, I doubt that these requests are causing your machine to lock-up. If they are, then something is very very wrong with your version of bind. I'd suspect that something else is causing your machine to behave this way. -- WWJD? JWRTFM Rot13 for email address: yvfgf @ ehqa.pbz |
|
|||
|
Brad Olin wrote:
> On Sat, 15 May 2004 19:07:06 -0400, Jeff Breitner <usenet@rudn.com> > wrote: > Ummm... Except that it is valid for those same verizon users to look up > a domain that his dns server is authoritative of. So you can't just > blindly drop them by address range and port. Well, unless the op has a > caching only dns server. If they are dial-up, then it's a safe bet that the radius auth session gave them Verizon DNS servers to use. They wouldn't be looking anything up at his nameservers, Verizon's nameservers would be doing that and passing the look-up along to their clients. I'd be really surprised if Verizon has those nameservers even remotely close in IP space to the radius assigned nameserver. Therefore, and as I indicated in my previous post, he could IPTABLE them with the only bad effects limited to the Verizon users within the IP space he blocked. And even then, it would be those that are running their own nameservers (assuming Verizon didn't have the radius assigned nameservers within the blocked space -- so he wouldn't want to block an entire /16). And really, the only people capable of trying to send bad update information would be those running their own nameservers. Unfortunately, the offending single user probably would shift around the assigned address, so IPTABLE would be an action of last resort. -- WWJD? JWRTFM Rot13 for email address: yvfgf @ ehqa.pbz |
|
|||
|
Brad Olin wrote:
> Not true if he is authoritative on one or more domains on the lookup of > those domains. To paraphrase the process... The process starts by the > client contacting the tld server, then based on that reply, the client > contacts the authoritative dns server(s). If his dns server is the > authoritative server, and he blocks the requesting IP, then the > domain(s) he serves would be broken for that IP/address range. You are correct, *if* this particular client is running their own nameserver. And that right there is part of the problem; some Verizon client is running their own nameserver and is running amok with update requests and he'd like to block it. > > Your assumption is fine if he has a caching only server. As well the scads of other clients that are just ordinary dial-up users. > > >>Unfortunately, the offending single user probably would shift around the >>assigned address, so IPTABLE would be an action of last resort. > > > True if it's a malicious intent. It could be just a poorly configured > box. I lean toward the latter. The only malicious update intent I've seen was toward Windows DNS machines that allowed people to poison the cache and do such nasties as point all requests to some porn site. -- WWJD? JWRTFM Rot13 for email address: yvfgf @ ehqa.pbz |
|
|||
|
Brad Olin wrote:
> I have setup several caching only dns servers for dhcp customers > (dsl/cable/dialup) because the provider's dns servers are so slow or > overloaded. I don't know that there is any foul with that if it's done > correctly. I do it myself with my cable connection at home. > My cable ISP is the same way. DHCP assigned nameservers are slower than tax refund checks. -- WWJD? JWRTFM Rot13 for email address: yvfgf @ ehqa.pbz |