Resolver
Mar 7, 2007I have my servers at ev1 and can I use any ISP's IP address to resolve?
View 6 RepliesI have my servers at ev1 and can I use any ISP's IP address to resolve?
View 6 RepliesWhat is the best resolver ips for webhosting services?
I used 4.2.2.1-3 on my servers now.
For the last several days on one of our dedicated servers, a AMD 3000+ 2GB RAM 2x80GB HDD 5TB traffic installed with CentOS 5.3 64-bit w/a GNU Linux kernel-2.6.18-128.1.10.el5, we use as a 'backup' to our master, we've been having resolver issues. If we attempt to ping, dig or nslookup ANY hostname, we get "unknown host domainexample.com". We can, however, ping IP addresses, including the external one for the server. (Pinging the server from another, not connected to the network returns 0% packet loss.)
Our package managers can't resolve hostnames, either. For instance, when we try to run yum, we get:
"Could not retrieve reponamehere [url]
[Errno 4] IOError: <urlopen error (-2, 'Name or service not known')>
Error: Cannot find a valid baseurl for repo: reponamehere"
We've already tried disabling several repos and it should be noted that when we use "wget" or "git" or any other package installer, we have the same resolving issues.
Our "/etc/resolv.conf" contains nameservers for the major upstream DC to our provider, which gave us permission to use them. We've tried two sets from them as well as those from our provider and from a DNS service provider. None have worked. (BTW, We have two servers with this provider and the other one is fine; the provider, though this is an unmanaged sever, is attempting to be very helpful and responsive. No problems there.)
We've tried rebooting the server and flushing IPtables as well as stopping it altogether. (Thankfully, we've set our DNS, SQL and httpd servers to run on boot) Nothing seems to resolve the issue (no pun intended).
However, we think this problem may have to do with having installed Suhosin on the server this past weekend since it seems to have started shortly after that but we're not sure. We are getting the following errors in from our httpd logs, though:
[Fri Jun 19 12:52:25 2009] [notice] Graceful restart requested, doing restart
[Fri Jun 19 12:52:25 2009] [error] (9)Bad file descriptor: apr_socket_accept: (client socket)
[Fri Jun 19 12:52:26 2009] [notice] Digest: generating secret for digest authentication ...
[Fri Jun 19 12:52:26 2009] [notice] Digest: done
PHP Warning: PHP Startup: Unable to load dynamic library '/usr/lib64/php/modules/suhosin.so' - /usr/lib64/php/modules/suhosin.so: undefined symbol: php_rfc1867_callback in Unknown on line 0
[Fri Jun 19 12:52:26 2009] [notice] mod_python: Creating 4 session mutexes based on 256 max processes and 0 max threads.
[Fri Jun 19 12:52:27 2009] [notice] Apache configured -- resuming normal operations
When we ran 'tcpdump', all it returned was the IP address for the server connected on several ports via SSH. Running 'netstat -an' shows the server's IP address, again connected on different ports, and the localhost address connected on a few. No other IP addresses--and we have about 10--are showing as connected.
Anyway, this is all we've been able to figure out. Anyone had this problem and solved it successfully?
Seems that the default installation of Bind on Windows plesk 11.x (i don't know if is also for Linux), configure bind by default as open resolver. Then we must fix the problem using this kb:
[URL]
I'm wondering if i'm doing a wrong installation, of this type of installation is by default. In the second case, is there any reason?