Database Testing

Jeremy Kister argus-01 at jeremykister.com
Thu Mar 4 14:43:44 EST 2004


On Wed Mar 3 16:26:58 EST 2004, Jeff Weisberg wrote:

> Service DB {
>       user:         kister
>       pass:        alyssa

great.  now i have to go change my password. ;)

Interestingly, when i did use the incorrect password to connect, the service
showed as down (as expected), but Argus did not show me the reason (ie: bad
credentials) -- even though i have 'showreason: yes' in the config.

> then I'd start up my web browser, and look at the 'debugging' page
> for the service and look at 'db::rbuffer' to figure out what mysql
> was returning. then I'd add the appropriate 'expect' to my config.

well, the query returns a bunch of stuff (18 key/value pairs, to be exact).
db::rbuffer shows:
mysql.example.net user 3306 60 host1-bin.002 276511 host23-relay-bin.001 30.
. .

which are only some of the values.  since db::rbuffer is displaying these
values in a different order than my perl is, i cooked up a quick patch to
arguscgi that expands the debug info when ". . ." is placed on the end of
the string, when you hover your mouse over the debug text (using MSIE, at
least).
http://www.jeremykister.com/jeremy/code/argus/arguscgi-3.3_RC-20040217.expand_debugging.patch

the entire db::rbuffer contains:
mysql.example.net user 3306 60 host1-bin.002 276511 host23-relay-bin.001
3096735 host1-bin.002 Yes Yes dns,isp  0  0 276511 3096735~x0A

Without having the key/value pairs to test against, and since it is safe to
assume that the values i am looking for will no doubtedly shift around for
one reason or another (such as one of the values being null unless there was
an error at some point), i am weary in using "expect: Yes Yes"... are my
concerns founded?


Jeremy Kister
www.jeremykister.com/jeremy/



More information about the Arguslist mailing list