Microsoft SQL Server - Partial Result: U

[timmeh]

Hi there,

I have an issue with the SQL Server template. If I process the same command at command-line then I successfully get the correct response.
When reviewing the cacti log file I can see the following:

SERVER: C:/cacti/scripts/ss_win_mssql.php ss_win_mssql IPADDRESS:1433 logflushtraffic USERNAME PASSWORD, output: U
WARNING: Result from SERVER not valid. Partial Result: U

Why is it that I get the right response at command-line but not when its running automatically?

Cheers,
Timmeh!

[lints]

I just saw this post...

Has the issue been resolved or is it still doing it?

[timmeh]

Still the same I'm afraid mate, any ideas?

Many thanks for the response.

[lints]

If you start the script server manually and run the script, does it still crap out?

Can you temporarily comment out the 'error_reporting(0);' line to generate more debug info?

What version of SQL Server are you polling?

[timmeh]

I plan to be connecting to 2005 upwards but currently testing with 2008.

Just so that you know I have had to modify the php due to the SQL DLL being the new version so the commands have changed.

If I run the script server at command line and then run the script the result is ok.

i.e

start > run
C:\cacti\: php\php.exe script_server.php
then run:
C:\cacti\scripts\ss_win_mssql.php ss_win_mssql IPADDRESS: logflushes

Yet if I look in the Cacti logs if comes back with result U

[lints]

Wow... I have NO idea then.

I'm running it on Linux and I've never seen any issues like that.

I know some people reported issues with Windows not liking the colon and port setting, but I fixed that issue at some point and re-posted the script. That said, you've torn that bit of logic out cause it doesn't apply to you.

What about turning on the error_reporting? Does it throw any errors or warnings?

[timmeh]

Hi,

Error Reporting is set to E_ALL
Running via the command line in script server no errors are shown just the correct results as expected.

When the script server runs automatically is there any real difference from when I run it as a command line?

Any way to show the actual error rather than just U?


Thanks again,
Timmeh

[lints]

Dang.. That's frustrating..

There shouldn't be anything different running the script server manually as opposed to running it automatically, except maybe the user that's running it.

There's no way that I'm aware of to have only certain data sources poll in "debug" mode. It's all or nothing and if your poller has lots of items, it can be a haystack of information to find a few needles.

Speaking of which, what's in your poller cache for these items?