-----Original Message----- From: Henrik Stoerner [mailto:henrik at hswn.dk] Sent: Friday, February 16, 2007 3:14 AM To: hobbit at hswn.dk Subject: Re: [hobbit] Changing the Test interval
On Thu, Feb 15, 2007 at 02:49:51PM -0600, Trent Melcher wrote:
Is there a way to change the interval for either the content or http tests to 1 minute instead of the default 5 as everything else?
Change the INTERVAL setting for the [bbnet] task in hobbitlaunch.cfg
I was hoping I didn't have to do this
Could it be possible in the next release to have the option of testing different network tests at different intervals based on test or server? I would still like to test pings and most other net tests at 5 minute intervals but content or http I would like to be able to specify the interval per server.
Trent
Note that you should not do this if you have lots of systems that you poll, since it's possible that it just won't be able to run all of the network tests once a minute. In that case, you should pick out the important ones that you do want to run every minute, flag them with a "NET:1minutepoll" and setup an additional [bbnet2] task that runs a script like this:
#!/bin/sh BBLOCATION="1minutepoll" export BBLOCATION bbtest-net --ping --checkresponse exit 0Regards, Henrik
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
On Fri, Feb 16, 2007 at 01:50:58PM -0600, Trent Melcher wrote:
Change the INTERVAL setting for the [bbnet] task in hobbitlaunch.cfg
I was hoping I didn't have to do this
Could it be possible in the next release to have the option of testing different network tests at different intervals based on test or server? I would still like to test pings and most other net tests at 5 minute intervals but content or http I would like to be able to specify the interval per server.
I'll keep it in mind. I will be re-inventing how network tests run; there are several things on the "wishlist":
- Different test schedules (your request)
- The ability to request immediate testing of some service (e.g. to see if it's recovered without having to wait for the next test cycle).
- Better dependencies, i.e. some form of network topology knowledge so we can better handle when a whole subnet disappears because of a router- or link-outage.
I have an idea for a network test "daemon" which is fed test-requests from a kind of "test scheduler". The test daemon just reports the raw result of the network test, but feeds it into an analysis module that has the detailed knowledge of eg network topology and other kinds of dependencies - e.g. this could handle clustered or load-balanced network services intelligently. The analysis module would then generate the final status messages, not unlike how the client data are processed by hobbitd_client before it becomes the actual status you see on the webpage.
Currently just an idea, and I'm probably rambling due to the fact that we had a major network outage yesterday and I was working until 3 AM ...
Regards, Henrik
- The ability to request immediate testing of some service (e.g. to see if it's recovered without having to wait for the next test cycle).
This would be really nice for two cases:
When a broken state is detected, it would be nice to be able to re-test manually and confirm. Avoid trying to fix a "blip."
After something is fixed, it would be nice to re-test immediately to confirm it really is fixed.
Scott
participants (3)
-
henrik＠hswn.dk
-
scott＠PacketPushers.com
-
tmelcher＠saionline.com