- Xymon - Xymon mailman web

can't get rid of old status with "*" in its name
by cgoyard＠cvf.fr 30 Aug '06

30 Aug '06

Hi, Someday, I made a mistake and sent a status column whose name is "tomcat_*". I then dropped the status with "bb drop". But each time I restart the hobbit server, the status pops up in purple state, again and again... I tried to manually delete most files that refer to this (history, rrd), either with the daemon running or not, but without any luck. Where does the hobbitd keeps its status when shutting down ? -- Charles Goyard - cgoyard at cvf.fr - (+33) 1 45 38 01 31

2 2

Renaming a host and keeping the history - lessons learned
by jonescr＠cisco.com 30 Aug '06

30 Aug '06

I started to type this email to describe a problem I was having, and in the process I figured it out. I thought I would share it for amusement and learning purposes :) I have several hosts that I had to rename. Their DNS and hostname names changed but IPs remained the same. Example: myhost-web-1 -> myhost-demoweb-1 myhost-app-3 -> myhost-demoapp-1 myhost-db-2 -> myhost-demodb-1 I wanted to keep the historical trending info, so I proceded as follows: 1. Stopped the Hobbit server 2. Made the changes to bb-hosts, simply replacing the hostnames 3. renamed data/*hist*/myhost-web-1.* to data/*hist*/myhost-demoweb-1 (did the same for the other hosts) 4. renamed data/*histlogs*/myhost-web-1 to data/*histlogs*/myhost-demoweb-1 (did the same for the other hosts) 5. renamed data/*hostdata*/myhost-web-1 to data/*hostdata*/myhost-demoweb-1 (did the same for the other hosts) 6. renamed data/*rrd*/myhost-web-1 to data/*rrd*/myhost-demoweb-1 (did the same for the other hosts) 7. Restarted Hobbit After the restart, the new hostnames appeared, but the only status I had was conn and ssh. After double-checking that I renamed the files properly and then wondering why I was getting no status updates it hit me. The hobbit *clients* had not been restarted! So Hobbit was polling them (I use /pulldata/), and they were reporting their old hostnames. :-) I restarted the clients on the hosts, and viola! They restarted with their proper hostnames and Hobbit could now recognize them. I am pleased to report that now everything is now as it should be, and the trending data is intact. Ways I could have figured this out quicker: 1. Thinking it all the way through from the beginning. I was smart enough to stop Hobbit before I began so that it would not create or overwrite the rrd and other files that I was renaming, but I didn't think to restart the clients. 2. I could have checked the ghosts report (hobbitd status) and noticed that the clients were reporting with their old names (I didn't notice this until after I fixed it) Hopefully this "problem" and resolution will prove helpful to someone ... even if it's only to help them chuckle at my newbie mistake :) -Charles

2 1

AW: Re: [hobbit] Hobbit 4.2.0 hobbitcgi.cfg
by bb4＠richter-it.net 30 Aug '06

30 Aug '06

Hi Henrik, >On Tue, Aug 22, 2006 at 02:19:16PM +0200, Torsten Richter wrote: > >> first of all I have to agree with all the others. >> This IS a really outstanding piece of software you've written. Thanks for >that. > >My pleasure. I am really happy that this version has turned out to be >fairly solid. > >> I found a small typo in the above mentioned config file. >> In hobbit-ghosts.sh you use "...hobbit-ghosts.cgi $CGI_GHOSTS_OPTS" but >> in hobbitcgi.cfg the last S is missing so it reads "CGI_GHOSTS_OPT=...". > >You've probably upgraded from a beta- or RC-version. This was fixed in >the final 4.2.0 release version, but upgrading won't replace an existing >configuration. No, I had version 4.1.2p1 up and running and fetched the archive from a Sourcforge mirror. Then I did a fresh install in a new directory. -rwxr-xr-x 1 hobbit users 190 Aug 22 12:18 hobbit-ghosts.sh ./bb 127.0.0.1 ping hobbitd 4.2.0 The mirror I chose was MESH solutions in Duesseldorf. Best regards Torsten

4 3

case sensitive bb-hosts?
by gjohnson＠trantor.org 29 Aug '06

29 Aug '06

Not sure if this has come up before but I ran into an issue that I would consider a bug...others may not as it's more of an annoyance. After setting up 4.2 and a BBWin machine I kept getting a ghost machine listed under the hobbitd montor. I had entered abc1.fqdn.test.com into the bbhosts file...I got my connection and http info...I checked the hobbitd monitor and saw a "ghost" ABC1...then I changed the entry to abc1...same output...once I changed the entry in bb-hosts to ABC1 (an exact match of the ghost in hobbitd), hobbit displayed the monitors as I would expect it to. The fact that bb-hosts is case sensitive came as a bit of a surprise. I would think you would want to run it through tolower...is there a reason it was set to be case sensitive? I can't think of a good reason for it to do so. =G=

2 2

[bbwin-users] CPU graphs in hobbit (not)
by gjohnson＠trantor.org 29 Aug '06

29 Aug '06

Etienne Grignon wrote: > Hello, > > 2006/8/29, Galen Johnson <gjohnson at trantor.org>: > >> Has anyone else run into any issues with the output from BBWin that is >> being fed to Hobbit not graphing? I installed BBWin .80 onto a Win2K3 >> box today and the cpu graphing on hobbit isn't working...I'm not where I >> can send any output but if needed I can try to give more feedback >> tomorrow... >> >> =G= >> > > What hobbit version are you using ? > > BBWin cpu report has the same format than the bbnt client so graphs > should not be broken with BBWin. Could you post your cpu report ? > > Do you have any image graph or is the image present with "nan" values ? > > Did you check your hobbit log for rrd errors ? > > Regards, > > Actually, sometime during the night it started graphing...so it looks like I spoke too soon. It would be kind of nice if the graph showed all processors and the totals as a combined or separate graph (the average on the front page but when you click to drill in the next page look something like a disk graph) but I think that would fall more onto the Hobbit side... =G=

1 0

how to adjust timeout value for HTTP/SSH tests ?
by jjj863＠gmail.com 29 Aug '06

29 Aug '06

hi, there, Once or twice a day, I have a server whose HTTP/SSH will go red then recovered within a minute. For a few times, it recovered at next run instead of within the same run. The somewhat sluggish response from HTTP/SSH servers could be due to the load on the system at the time. For this particular server, I just want to know the service is responding and not all so concerned about its performance at this point. The actual status message says the server times out at 10.79s. 1) any way I can set this timeout value to 30s or 60s 2) it seems there's a 2nd probe for each test within a test run. Can it be set to go red only failed two probes in a row? I assume the 2nd probe using longer timeout. red Mon Aug 28 18:10:07 2006: Server timeout &red http://server01/index.php <http://callisto/wiki/index.php> - Server timeout Seconds: 10.79

2 1

OS info missing on info page
by Dirk.Kastens＠uni-osnabrueck.de 29 Aug '06

29 Aug '06

Hi, the line with the OS information is missing for all clients that I updated to the 4.2 version (including the pre-release). The old clients that gather their information in a file called client/tmp/msg.txt are correctly displayed, while those with the file client/tmp/msg.<fqdn>.txt are not. The information in these files is complete, for example [uname] Linux sanodxx.yy.unizz.de 2.4.20-20.8smp i686 [osversion] Red Hat Linux release 8.0 (Psyche) but the information is not being displayed by the server. Any ideas? Regards, Dirk Kastens

1 0

trends question
by sladewig＠bankinfo.com 29 Aug '06

29 Aug '06

Greetings. On my trends pages I am not seeing all of the trends. Specifically iostat. I see the clients running iostat but nothing is being gathered. In hobbitserver.cfg I have TEST2RRD and GRAPHS which contain iostat. How can I get these to show up? thanks, steve

1 0

AIX client question -- data overflows Hobbit server
by greg.hubbard＠eds.com 29 Aug '06

29 Aug '06

I am monitoring some special purpose AIX systems, and every once in a while the client sends more data than the server can handle. It looks like the cause might be the routing table dump, which can be quite large on these systems. Evidently these systems dynamically add a lot of host-specific route entries. I don't recall anything on the Hobbit server that checks the client routing table, so is it going to be a problem if I simply disable this part of the client data gathering? GLH

2 1

[bug?] disable/enable does not set the color back
by cgoyard＠cvf.fr 29 Aug '06

29 Aug '06

Hi, When sending a long-term status (such as "status+30h"), if you disable/enable the status, it keeps blue until hobbitd gets a new one. Since it's ok for status whithout lifetime, it may be a misfeature for others. (I use +30h for inventory checking and some purely informational data, so refreshing the status once a day is ok for that matter). (i'll try to include this information into the info column in the future). Regards, -- Charles Goyard - cgoyard at cvf.fr - (+33) 1 45 38 01 31

1 0