Custom HTTP monitor that checks for existence of an image

I'd say you've taken a good approach. I'm not sure you actually need the Host or Connection parts myself and you may be making things overly complex with that. I'd try without and lose those if they are not required.

Hi Steve,

 

Thanks for your reply. I also was wondering if I had more in there than I needed. I have this monitor applied to a test pool of servers so the good thing is I'm able to alter it on the fly and see the results. I tried a couple of variations including: GET /branding/_1_1/logo_uchc.gif HTTP/1.1\r\n\r\n\ and GET /branding/_1_1/logo_uchc.gif HTTP/1.1\r\nConnection: Close\r\n\r\n but the monitor then fails in both cases. I'm thinking I might just need to leave it the way it is since I don't fully grasp the syntax.

 

Rob

 

I would have thought GET /branding/_1_1/logo_uchc.gif HTTP/1.1\r\n would have been enough. You should be able to test what works by telnetting to the server:port from the device and manually inputting the request string. The only thing that might be different between that and the real monitor is the \r\n, v10 adds it automatically, v9 and v11 don't I believe.

It looks like going directly against the server, it will only accept: GET /branding/_1_1/logo_uchc.gif

 

Everything elese comes back with "HTTP Error 400. The request hostname is invalid." or " HTTP Error 400. The request is badly formed.".

 

Thanks,

 

Rob

 

OK, so your monitor could probably reflect that? The only thing I'm not sure of is whether you need the \r\n?

So after a little more testing, it seems to be happy with: GET /branding/_1_1/logo_uchc.gif HTTP/1.0 \r\n\r\n

 

The server directly responds to: GET /branding/_1_1/logo_uchc.gif HTTP/1.0 so I think HTTP/1.1 was the issue even though it appeared to work through the LTM.

 

Thanks for the help Steve!

 

Rob

 

You're welcome, glad it's working.

Hi Rob,

 

 

Of course LTM supports HTTP1.1 and if your server supports it, why no to use it - even for the monitor. Based on your info, you did two mistakes when testing/configuring the monitors:

 

 

1. MONITOR SYNTAX

 

- The original monitor was correct and, as you said, it was working (GET /branding/_1_1/logo_uchc.gif HTTP/1.1\r\nHost: host.domain.com\r\nConnection: Close\r\n\r\n)

 

- The other two variations were missing the HOST header, which is mandatory for HTTP1.1 protocol thus they didn't work as the server was expecting the host header in HTTP requests

 

- The value of the host is not significant unless the server expects specific host value or it's configured to return different content based on the host values; You can use virtual domain name, IP address or the real server name; I believe 'blank' value or space might also work, but I'm not 100% sure

 

- The 'Connection: Close' header is not mandatory, but it's good to use it so you don't have too many opened TCP connections to the server if your monitor sends the health checks too often

 

 

The following two SEND strings should both work for you:

 

 

GET /branding/_1_1/logo_uchc.gif HTTP/1.1\r\nHost: host.domain.com\r\nConnection: Close\r\n\r\n

 

GET /branding/_1_1/logo_uchc.gif HTTP/1.1\r\nHost: host.domain.com\r\n\r\n

 

 

2. TELNET TEST

 

- Note that when trying to get to the server directly by telnet and then typing the GET request is different from the send strings specified in the LTM monitor:

 

- \r\n sequence is represented by hitting the ENTER key in the CLI, so if you want to test exactly the same string specified in the LTM monitor from the CLI, you need to separate each request line or header with the enter key; LAST line must be followed by blank line, which is DOUBLE ENTER in CLI or \r\n\r\n sequence in the LTM monitor:

 

 

telnet {server_name} 80{ENTER}

 

GET /branding/_1_1/logo_uchc.gif HTTP/1.1{ENTER}

 

Host: host.domain.com{ENTER}

 

{ENTER}

 

 

or

 

 

telnet {server_name} 80{ENTER}

 

GET /branding/_1_1/logo_uchc.gif HTTP/1.1{ENTER}

 

Host: host.domain.com{ENTER}

 

Connection: Close{ENTER}

 

{ENTER}

 

 

Peter, thanks for the additional detail and formatting information!

Peter,

 

Thanks so much for chiming in and explaining the syntax! Both telnet examples you provided returned the image file and I'll be changing my monitor back to what I had originally. Thanks again!

 

Rob