I have a Win 2003 Server that its only duty wis a web server. It is running iis 6, php4, mysql4 and WinFTPserver. There is nothing else running on the pc at all.
After usually 8-12 hours of running, the server will not function anymore becuase it will not connect to the MYSQL database. I looked in to it further and found that if I open up Internet Explorer, I cannot browse to any web pages, even if I put the direct ip address into the browser. This is why my web site will not connect to my SQL database, because it is looking for "Localhost" in the code.
Network configurations are: THis server is on a domain network with a main domain server (not this PC). There is a gateway router that has a NAT routing configuration to forward the IP traffic directly to this PC. The router has three different static IP addresses, one being the main ip for network internet traffic, one for the web site in question, and another that is not used.
Went throught testing procedures and found the following: 1. If I restart iis, no change. 2. I can access the MYSQL database remotely, by using the computers direct ip address. 3. IIS is funtioning properly from a remote computer, I can access any static .HTML or .php pages that do not need to connect to the MYSQL database. 4. On the local computer, I can browse to other computers on my network such as //SERVER without a problem. 5. I thought it was a hardware problem so I built a new web server with windows xp, and the same configuration. I copied all of the web pages over and served pages for a day. SAME PROBLEM!! 6. If I restart the server, the problem disapears for 8-12 hours and then reappears.
What Changed?? THis web site was running fine for over a year. My server crashed and I had to rebuild it. I recompiled all of the programs such as php, iis, etc. I may have used different versions, but for the most part. It was the same. I did not have the other server connected inside my domain. It was connected directly to the internet with the ip address that my domain name resolves to. Since it is now inside the domain network, I have the router set up with NAT in order to translate the 192.168.0.** name into the 74.99.99.99 domain name ip.
I think thats it. I need help bad here. This is screwing up my world big time.
Thanks for the help but niether of those provided a solution. The problem that I am having is intermidant. It happens after usually about 6-10 hours of up time. When I come in to the office in the morning, it is down. I completely restart the server and it fixes the problem for a while then it will reoccur 6-10 hours later.
Assuming the web server has a static IP configured. ---Static IP configured....Yes
And you said it is now a part of a domain. I am assuming it is an actual member of the domain. ---Yes, actual member of domain
In the IP configuration, the DNS settings should be pointing to the domain controller only. ----Domain controller is primary, ISP DNS is secondary
The domain controller DNS is probably set to Active Directory Integrated mode. ----True DNS forwarders on the DC are configured for your ISP's DNS servers. ----True
It did not work. Checked it this morning, same problem.
Something else I did notice, dont know if it has anything to do with it. When I tried to remote login, it said that the RPC Server was unavailable. Probably the same problem with no DNS functionality but ??
The RPC Server unavailable is troubling. That is a problem with your domain controller. It's possible that is the cause of your DNS problems. I would fix the rpc problem before spending any more time on this issue.
Oh, also. THis computer cannot find the RPC server because it looks for the ip address and cannot resolve it. I cannot connect to 192.168.0.254, which is the server at all.
OK, this is a strange one... Let's check a couple of things: 1. When you have the problem, can you ping the DC by name? by IP address? 2. Any odd entries in the event log regarding security, kerberos, authentication? 3. On the DC, in DNS, the server is configured to receive dynamic updates. Secured or not? 4. Is there another nic in the webserver that can be used for troubleshooting? Or one available to swap out? The next setp we're going to have to try is to remove the webserver from the domain, reboot, then add it back to the domain.
1. When you have the problem, can you ping the DC by name? by IP address? Yes by name, no by ip address. 2. Any odd entries in the event log regarding security, kerberos, authentication? Only thing is the error "The browser service failed to retrieve the backup list too many times on transport\device\......the backup browser is stopping." This error occurs when it is working or not.
3. On the DC, in DNS, the server is configured to receive dynamic updates. Secured or not? Yes it is configured to recieve dynamic updates. I did find this error in the log that occurs every day. It has been in the system log daily for over a year so I dont know if dynamic updates is working properly.
Dynamic registration or deletion of one or more DNS records associated with DNS domain 'ForestDnsZones.mge-us.com.' failed. These records are used by other computers to locate this server as a domain controller (if the specified domain is an Active Directory domain) or as an LDAP server (if the specified domain is an application partition).
4. Is there another nic in the webserver that can be used for troubleshooting? Or one available to swap out? No and Yes but this cannot be the problem becuase after the initial crash of my old web server, I built a temporary web server with a windows XP pc. I built the server with this PC from scratch with a clean hard drive. After it was complete, I had the problem I am having now with that pc. I initially thought it was a problem with the NIC, so I built a new server with Server2003 OS, a clean install also and the same problem occured.
The next setp we're going to have to try is to remove the webserver from the domain, reboot, then add it back to the domain. Due to what I said in the above comment, I think this has already been tested since the first web server and the second web server both were added to the domain with a clean install and they both failed the same.
So we can be sure that the problem is not hardware related since two computers that were both clean installed had the same failure.
You can ping the DC by name, but not by IP address? I don't see how that is possible. Does the DC have multiple nics and/or multiple IP addresses? When you successfully ping by name, does it resolve to the same IP address?
Well its hard for me to test it now because this is a production website so when it goes down, I have to restart it right away but I do recall that this was the case before. I didnt go to the //DCserver directly but I know that my file folders that were shared were still accessible by.
I also cannot recreate the problem on demand either since it only occurs every 8-12 hours so since I restarted at 730 am today, it wont go down till around 6 or so most likely.
So to clarify what I did know is that //fileserver was still accessible when the problem was occuring but I did not test the DC server with the same name protocol. THis was shown in network drives that the web server could access.
I didnt understand this either. It does not seem possible.
Ok, got some new info for you. Computer actually went down today at lunch. Starting to get a better handle on the problem now.
So when the problem was occuring I found the following. 1. I COULD ping any ip address or website from the command prompt with no problem. 2. I COULD NOT navigate to any web page in internet explorer or windows explorer. No matter if I used direct IP addresses or not. Whe I put direct IP addresses into Win explorer, it would try to bring up the Microsoft default search engine. Saying "Page cannot be displayed". I could not navigate to any //localcomputernames or //localhost either.
So DNS is definitly not the problem since CMD prompt Ping resolved google.com, msn.com etc. into IP addresses. THis would lead me to believe that the local PC is not allowing more connections to be opened up since it can ping these sites but not connect to them.
Testing things further found out the following. -Does not seem to be memory leaks, running program that is monitoring system functions and all seem ok. - Set up a windows xp server with the same programs and configurations running. It did not have NAT ip translation or any traffic going to the PC at all. Still failed the same way after four hours (as the 2003 server) . Could ping but not connect to any ip addresses. -When the problem occured I tried : release/renew ip...did nothing Tried to connect to a mapped network drive. Did not work -All of the other regular pc's in the office are working fine, most are domain members. No problems noticed with them.
So I took the computer off of the domain "physically" without making any changes. I hooked ethernet cable right to the WAN modem.
So there was no physical connection to any computers in the domain....I set the ip address to the WAN address, used the dns from my ISP...etc etc. and same problem occured about 5 hours later.
So this leads me to believe that the problem has no connection to the domain or DC at all.
Comparing the two simlar computers that I can create the issue on, here is what I know so far. Same problem with either IIS 5 or IIS 6 Same problem with PHP5 or PHP4 Both running MYSQL 4.1 Both running WINftp server Both have OS installed from scratch Same Problem with WINXP or WINServer2003 Both are members of domain but it doesnt matter if it is actually connected to domain.
Today, the problem occurred quicker than normal. First time today 3hr45min, second 5hrs15min, third time, 4hrs30min. Traffic was heavy during the day, light at night but it didnt have any effect on the amount of time before site crashed.
OK, so we've eliminated the following: Domain issues Basic networking DNS Operating systems IIS PHP Physical connectivity Hardware
The way I see it, there are only 2 things left: MySQL or the specific configuration/settings you are using. Unfortunately MySQL is not my area of expertise, so I'm not sure how much help I can be. I've got to admit, this has me stumped. I haven't given up, just can't think of anything right now. Will talk to some of my programmer friends and get back to you.
So I dont know if you are still watching this but I had some strange findings over the weekend.
I was able to eliminate the problem after I removed the computer from the domain physically. I took my web server that was still a member of my domain and hooked it up directly to the modem. After the first night, the computer showed the same problem as it did before. I did a reboot on friday in the morning and ever since, it functioned perfectly. So it definitly has something to do with the domain but I dont understand why when I removed it from the domiain physically, it failed the first day but then the next two days, it worked fine.
Ok, scratch that last idea that I posted. It went down even not hooked up into the domain. It took a lot longer this time (2.5 days) instead of the usual 6 hours. ??????
Sorry for the delay, was out of town... I spoke to some programmers I know, they had no ideas that we haven't already tried. I hate to admit it, but I thiink I'm stumped. Have you posted this problem anywhere else? I'm a member of sever other technical forums, someone there may know of something else we haven't thought of. Try posting here: http://www.tek-tips.com/threadminder.cfm?pid=931 I will continue to monitor both sites and will assist any way I can.
