1 |
On Oct 18, 2011 12:14 PM, "Norman Rieß" <norman@×××××××××.org> wrote: |
2 |
> |
3 |
> -----BEGIN PGP SIGNED MESSAGE----- |
4 |
> Hash: SHA1 |
5 |
> |
6 |
> On 10/17/11 20:06, Pandu Poluan wrote: |
7 |
> > |
8 |
> > On Oct 17, 2011 6:44 PM, "Norman Rieß" <norman@×××××××××.org |
9 |
> > <mailto:norman@×××××××××.org>> wrote: |
10 |
> >> |
11 |
> >> |
12 |
> >> Hello, |
13 |
> >> |
14 |
> >> sorry to interrupt this thread, but this probably means, you did not |
15 |
> >> perform any kernel updates on that machine for over two years and |
16 |
> >> therefore the system is vulnarable to some kernel bugs which where |
17 |
> >> discovered during this time. On a DNS machine a privilege escalation |
18 |
bug |
19 |
> >> is even more severe. I strongly recommend to secure this machine. |
20 |
> > |
21 |
> > That depends on what Kai meant with "uptime". Maybe he meant the VMs |
22 |
> > (he's using Xen, after all) never needs a restart, but the BIND service |
23 |
> > still gets regular update and the consequent service-restart. |
24 |
> > |
25 |
> |
26 |
> Every Xen VM is running its own kernel and needs to be restarted or |
27 |
> kexec'ed when this kernel is updated. If this is not the case, the VM is |
28 |
> vulnerable to kernel bugs just as any other physical system, even if the |
29 |
> host on which the VM is running is secure. |
30 |
> I assume BIND is updated and restarted as needed, but that is not enough. |
31 |
|
32 |
Does it matter if the DNS server is behind a firewall that allows only |
33 |
TCP+UDP traffic to port 53? |
34 |
|
35 |
Rgds, |