September 2007 - Xymon - Xymon mailman web

Hobbit Server H.A. idea: hobbit client rework
by tj_yang＠hotmail.com 07 Sep '07

07 Sep '07

I am interested to see Hobbit server to have High Availability soluation. There were some discussions on the list. Following is another idea. The current design does allow hobbit client to send message to two hb server at the same time. See following, hbclient> grep ^BBDISP hobbitclient.cfg BBDISP="0.0.0.0" # IP address of the Hobbit server BBDISPLAYS="hbs.test.com hbs2.test.com" # IP of multiple Hobbit servers. BBDISP must be "0.0.0.0". hbclient> But this duplication can impose network traffic penalty on a site with large hb clients installation. The data from hbs to hbs2 can be mirrored using DRBD(hb server on Linux) or AVS(Hobbit server on Soalris). 1. Rework hb client send to hbs2.test.com in BBDISPLAYS list ONLY if hbs.test.com is not reachable ? 2. or adding a new variable BBDISPBK when BBDISP or BBDISPLAYS servers are not reachable ? Comments ? T.J. Yang _________________________________________________________________ Get a FREE small business Web site and more from Microsoft� Office Live! http://clk.atdmt.com/MRT/go/aub0930003811mrt/direct/01/

3 4

FW: [hobbit] scripting help
by Jason.Chambers＠geosoft.com 07 Sep '07

07 Sep '07

This email didn't make it onto the list, so I thought I forward it here. Jason Chambers I.T. Helpdesk Associate Geosoft Inc. 85 Richmond St W - Toronto, ON, CA - M5H 2C9 - Tel: 416.369.0111 ext 344 - Fax: 416.369.9599 -----Original Message----- From: Jason Chambers Sent: September-07-07 9:05 AM To: 'Buchan Milne' Subject: RE: [hobbit] scripting help Hi Buchan, I didn't see it until you mentioned it. Email got blocked by our spam filter for some silly reason. ;) It looks like bbhostgrep does not like : entries in the 4.2.0 release. I think I might need to use bbhostshow and parse it manually, as buggy that might be :\ unless I am missing something. Jason Chambers I.T. Helpdesk Associate Geosoft Inc. 85 Richmond St W - Toronto, ON, CA - M5H 2C9 - Tel: 416.369.0111 ext 344 - Fax: 416.369.9599 -----Original Message----- From: Buchan Milne [mailto:bgmilne at staff.telkomsa.net] Sent: September-07-07 8:51 AM To: hobbit at hswn.dk Cc: Jason Chambers Subject: Re: [hobbit] scripting help On Friday 07 September 2007 14:35:54 Jason Chambers wrote: > Hi Iain, > > I really don't want to install an smnp client on each windows machine, > which may or may not affect the performance of a mission critical servers, > to do the polling of data, when I can actually get the hobbit server to > fire a script. > > Also, what or how I am executing this script is not the question I am > placing here. I'm just wondering how I can read parameters after the test > name in bb-hosts (my planned version of blades:(blade port number):(ip > address)) that is very similar to the setup of how Henrik has badtest:x:x:x Did you even look at the example script I sent? Did you reply saying what it didn't do that you need? Or, are you looking for someone to teach you how to read and write scripts (in which case I think you're looking in the wrong place).

1 0

scripting help
by Jason.Chambers＠geosoft.com 07 Sep '07

07 Sep '07

Hey guys, I want to make a script that monitors individual blades in our H-Series bladecenter via SNMP. What I want to do is put in the bb-hosts file blade:# where # would indicate the blade slot number that would be used on snmpget against the bladecenter to get the information on that blade. Any ideas on how I could go about this? Jason Chambers I.T. Helpdesk Associate [cid:image001.jpg at 01C7F08A.8AF7FED0] <http://www.geosoft.com/> 85 Richmond St W<http://maps.google.com/maps?f=q&hl=en&q=geosoft&sll=43.650833,-79.382853&ss…> - Toronto, ON, CA - M5H 2C9 - Tel: 416.369.0111<skype:+14163690111> ext 344 - Fax: 416.369.9599

7 16

bb-hosts: Same server multiple times: "noconn" use adviseable?
by haertig＠avaya.com 07 Sep '07

07 Sep '07

I have hosts that appear multiple times in bb-hosts. I pick one instance and give that the "prefer" tag and the real IP address. Other (secondary) instances have an IP of 0.0.0.0 and do not have the "prefer" tag. Is it adviseable to also put "noconn" on each of the secondary instances? I currently have done that, but would like to remove it if it's a useless thing. But then, I would want to leave "noconn" in place if that prevents Hobbit from excessively pinging a server with multiple instances. Is "noconn" a NOP in my scenerio, or is it actually doing something useful for me? I wouldn't know where to check to see if it's preventing excessive pings or not. Thanks!

3 2

Blue, Green, Red -- Disable Individual Tests within a Test
by jkwade＠futurefrontiers.com 06 Sep '07

06 Sep '07

Hello, So, we are looking at monitoring MQ and if the Queue Manager is up. However, in development, they may want to keep some of the Queue Managers down. Of course the Test show's red when a Queue Manager is down. What I would like to know, is it possible for the overall test to indicate Green or Yellow, and monitor the other Queue Managers while allowing an individual Queue Manager to be blue. I hope I described this correctly. So, Basically, I have a custom hobbit script that checks multiple items. If an item breaks, it goes red. What I want to do is be able to disable the single broke item (using blue), while the other tests continue monitoring. If I click on the MQ test, it would show green, but with a single item being blue because I disabled it. Has anyone done something similar to this in another test. By the way, I have this same problem in the filesystem test. On a Database Server, I may have a single file-system that is 98% full and they are working on coming up with a solution to reduce the space used. However, I don't want to disable all the other filesystem tests while they are working on the other. It would be nice to be able to disable the check of the individual filesystem turning it blue, but the other filesystems are checked and would re-alert us if a different filesystem goes red. James

1 0

Trying to find a sybase monitoring client that works with ASE 15.0.1 in Hobbit
by Anne-Marie.Lidzen＠vattenfall.com 06 Sep '07

06 Sep '07

Hi! On the site for BigBrother (link from hobbit) i found a new version of bb-sybase_ase1.2.sh. This script calls on a sql-script called sp__dbmon.sql. The problem is that I can´t find a new version of this sql-script. The old version (came with bb-sybase_ase.1.1.sh) of sp__dbmon doesn´t work with ASE 15.0.1. It shows wrong values for usage in Size (MB) Free (MB) Used (%) propably because of changes in structure for envolved system tables. Do anyone know where to find a newer version of sp__dbmon.sql or an other script to monitor sybase in Hobbit? Regards Ann-Marie

1 0

false Windows reboot alerts
by lebarber＠gmail.com 06 Sep '07

06 Sep '07

Has anybody else had problems with Windows machines showing a false reboot indicator in the CPU tests after about 8,000,000 or 16,000,000 seconds of uptime? I know you're not supposed to run Windows that long without a reboot ... Thanks, Larry Barber

4 3

Stupid sound question - "DNS Error"
by gumby3203＠gmail.com 06 Sep '07

06 Sep '07

It sounds like an obvious answer, but what does the http/https monitoring error "DNS Error" mean, exactly? Specifically, I have several https checks defined in Hobbit, and a good number of them get the "DNS Error" message multiple times a day; roughly on the order of every 1-3 hours. I have manually verified that DNS is working for these hosts. It might help me troubleshoot this if I know exactly how Hobbit comes to this determination, and if anything else other than DNS could cause it to appear.

5 9

issue with test time execution
by emarganne＠be.tiauto.com 06 Sep '07

06 Sep '07

Hello, Today I noticed something strange on my Hobbit monitoring system. Indeed we experienced an outage on our network causing a high number of hosts to be unreachable and therefore causing the network tests to be long. It took 1285 seconds instead of maximum 300. It looked that that thing leads some misinterpretations of the BBClient data coming from the monitored servers. All data coming out of the clients have been considered as valid even if they were generated something like 4 hours ago. Another thing is that even the network tests generated by the server have been green for a moment even they were not ok all (those parallel tests have been performed in such conditions that I am sure of what I am saying). This failure can cause some false positive if people are not watching attentively the test hours. If there are any ways to get rid of that bug... Any help or explanation is welcome. Etienne Marganne. DISCLAIMER: The information contained in this transmission may contain privileged and confidential information. It is intended only for the use of the person(s) named above. If you are Not the intended recipient, you are hereby notified that any review, dissemination, distribution or duplication of this communication is strictly prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.

1 0

Acknowledgements & History
by s_aiello＠comcast.net 06 Sep '07

06 Sep '07

All, I was curious if there was a reason why when an alert is acknowledged, why that acknowledgement is not displayed in History. In my mind an acknowledgement is a state change, i.e. Red -> to Red-acked. This would also help when performing audits on help-desk responsiveness to alerts in Hobbit. It would also help, when an alert acknowledgement expires. presently my team treats the alert as a new incident. It would be very helpful if they could check the history and see that the alert was previously acknowledged and what the ack-blurb was. Or are acks more of a BB legacy thing, and I should really be using disabled/maint of tests ? Thoughts ? ~Steve

2 2