Http 503 error

[swest]

I am running IIS Application server and about every day I get an HTTP 503 error, my website and my mobile application is down. After I reboot the server both works. Anyone else experiencing this?

[mikeallenbrown]

Re: Http 503 error

It would only a guess what might it be. Can you turn on the logging? At least then when the server goes down you'll have an error to track down.

[swest]

Re: Http 503 error

Here is the Failed Trace Log.failedtracelog.txt

[mikeallenbrown]

Re: Http 503 error

Yeah ...that's over 6000 lines.

[swest]

Re: Http 503 error

Alpha is saying it is a Denial Of Service Attack happening. I dont think it is, with a Denial Of Service Attack rebooting the server shouldnt fix the issue. Is Alpha saying they dont know how to fix their software?

[kkfin]

Re: Http 503 error

I thing the access log would tell much more what is happening in your site. There you can see if your server in under hammering or somebody using exceptional urls against your server.

[swest]

Re: Http 503 error

Where is the access log located at?

[kkfin]

Re: Http 503 error

:)

[swest]

Re: Http 503 error

So where is the access log located at?

[jgrannis]

Re: Http 503 error

The access logs (and you may find some other useful logs) are probably on the IIS server under something like:
C:\inetpub\logs (probably in the same general location as wwwroot)

Under logs, the access log should be found under: \LogFiles\W3SVC1

If it were an DoS and the attack brought the server down, whoever it was, probably stopped sending requests. So when you restart the server, it may restart without a lot of traffic and with no problem.

[swest]

Re: Http 503 error

I dont see huge amounts of traffic in the log files under \Logfiles\W3SVC1. I have decent size server on AWS so I would think it would handle a good amount of traffic before crashing. Any other thoughts?

[jgrannis]

Re: Http 503 error

Do you rely on any free-tier service in the global architecture? For instance, you have a good sized, not free-tier EC2 instance for IIS and the app but the app calls for resources from a free-tiers S3 bucket?
Or maybe the AWS Route 53 service that is configured "incorrectly". For instance, the "health check" configuration used to switch to another server is the server starts to slow down is a bit too sensitive, and perhaps, has no configuration to send the requests to a 2nd server. But I don't know if Route 53 would answer with a 503 or 504.

So many possible root causes that even if I had direct access to the IIS server to look, it could take a little while to set up additional tracing to find the source of the problem.
But I do love those types of problems. Obsessively tracking them down until I find it are always personally the best methods for me to learn.

EDIT: Thinking now that an unavailable or throttled down S3 bucket, would cause more of a timeout situation (504) rather than 503. I think the 503 would indicate an unavailable web server or a status sent by a gateway "in front" of IIS rather than anything slowed down or unavailable "behind" the IIS server. If you can't find and 503 entries in the IIS logs then it is very likely an element of the architecture in front of IIS that is responding with that HTTP code.

[jgrannis]

Re: Http 503 error

The failedtracelog.txt concerns one request to:

http://www.petvideoverify.com:80/con...Cms_Config.asp

.asp?

That specific request in the failed trace log triggers off a 401 status code and then the IIS configuration is set up to send back a HTTP code of 302 (but not before wasting a few resources looking for that .asp in the appserver's file cache). I don't believe that failedtracelog.txt contains any requests involving a 503 response from the server.
It is probably unrelated but you would probably want to look into solutions that prevent .asp, .php, etc., etc. from even getting handled by IIS or least by that particular hosted app on IIS.