Site problems

Need some advice and help

[agerhart]

I have a problem, and I am not sure what the hell is going on.

I just a friend, and old client, call me saying that she can't get to her site, and that a customer had come in and told her that he couldn't get there either.

I then went to the site, and got there no problem. I then typed in the URL, including the www, and couldn't get to the site. Then a minute or two later, I went back and checked it again, and I could get to the site.

I called her back and told her that I could access the site, and to check if she could get there now. Still couldn't get there. I asked a co-worker if they could get to the site, and they could access it.

I pinged the site.....everything seems okay.

The person who can't reach the site is in a different state, using a cable modem.............does anyone have any ideas what is going on?

[crash]

"I then went to the site, and got there no problem. I then typed in the URL, including the www, and couldn't get to the site"

sounds like a server issue.. but could be a router problem, jumping up and down. Keep an eye on it - and if it persists much longer contact the hosting company.

are there any other domains that you know of on the server? are they having the same problems?

[agerhart]

We already checked some of the other sites on the server and none of them are having any problems. I also had her go to a few of the sites that run on this server, and she got to them with no problem.

[crash]

1) is the site new? as in brand new? if so, could just not be propigated yet
*I have had instances where a new domain would appear on one computer (using dsl) but not on the one sitting right next to it (a different IP) doing tracerts proved that the two computers were using different routes (via same dsl co) to access the site - one route was not updated while the other was.

2) is the site setup properly? I ask due to it not showing up for www.site but showing for //site

[agerhart]

The site is over a year old, and has been working fine for the entire time.

[crash]

hmmm..

please understand I am guessing here since I only have what you posted to go by..

have you added any new sites to the server? this could have jacked with the mapping.. (virtual hosted right?)

or

there is a down router somewhere, in that case all you can do is wait for it to be fixed.. in my experience this happens rather quickly

or.. hmm.. can you email me the link? I will see if I can get it.

[Tapolyai]

It could be the following:

ISP line/bandwidth
You can verify this by running a tracert to it periodically, say once an hour. There are tools out there that can agregate several traceroute records and tell you the delays even the route changes.

DNS
If the ISP was messing with the DNS and sent some new records or resent them this could possibly cause problems but not likely.

Server itself
You can detect this unless you monitor each port with a tool, or sit on top of the monitor. You might consider an SNMP monitoring tool or some log server tool that monitors the health of the box.

client ISP
Can't do didly about this. If this is a big cheese then you can ping/traceroute their public IP for a while and watch dropped packets, and route changes, but you might just get banned. It won't tell you anything about the connection to the server UNLESS you do this tracing from the server.

client DNS
See above. You can run dig, and find out if their records were updated recently. Can ping the DNS servers, and see if they are acting up, but again can do didly.

Good luck.

[agerhart]

We have added many sites to the co-located server over the past year. If this was the problem, how would I be able to reach the site, and not her? Wouldn't it be down for both of us?

[agerhart]

Thanks to you both for the input.........hmmm, I guess I have some work to do to figure this out.

[crash]

I just emailed you back.

"We have added many sites to the co-located server over the past year. If this was the problem, how would I be able to reach the site, and not her? Wouldn't it be down for both of us? "

something could have messed up just that one record... it's not common but it does happen, took us forever to figure it out when it happened to us, fixing the record solved the problem...

For the record - I cannot get the site either, not via browser or tracert.

[mivox]

A few days ago Macguru was totally unable to reach my site on 4 different ISPs... I didn't receive any reports from anyone else unable to access it, and I could reach it just fine.

The next afternoon he was able to reach the site normally.

How long has the client been unable to reach her site? Has anyone else had problems?

Try to soothingly and patiently explain the idea of "temporary network problems" to her if it's only been a day or two. Also, give her instructions for doing a treaceroute from her location.

[agerhart]

I think that the problems just started today.

Thanks for the advice everyone. I think I will call her and do like Mivox just said.

Thanks again! ;)

[Brett_Tabke]

You have telnet on the box Andrew? Drop to a shell and see if you can ping HER computer. Then to a trace route from the box to her computer. See if there are any router problems.

[eeldivad]

this might fix the prob, but might not. i've had problems like that before that doesn't make sense cuz everything is setup correctly on my side and client side. so i guess its something in between.

you said that your client could access other sites you host fine, but it was just one site that had problems. well, try this cuz it worked for me. simply create another site, directory, permissions, and all. and then transfer all the files and settings from the "bad" site to the new one. hope this helps. make sure that u use different name for the site, and even the parent directory name should be different. if that works, then if u need to, change the names back to original.

i got a question now: how do u fix routing problems? how do u check the routing tables?

Thanks

[TallTroll]

I had a similar problem with a client site a while ago. The site was sometimes available, sometimes not. There was no apparent pattern as to who could/could not see the site, connection type or location etc. We finally figured out that the co-lo server providing 2ndary DNS had a copy of the primary zone file on it, and that was occasionally screwing everything up, making the site inaccessible, but only sometimes

Assuming that this is the problem, the differing routes used by different access methods may account for the wierd access errors

Worth a shot