August 2014 - Xymon - Xymon mailman web

CPU USAGE DISPLAY (CPU Statistics)
by feld＠feld.me 06 Aug '14

06 Aug '14

On Jul 23, 2014, at 10:51, Schwab, Jos <Jos.Schwab at newellco.com> wrote: > Hello everyone, > I am pretty new to the Linux world. Getting to the point quickly; xymon apparently uses the top display for the CPU display. When I see a CPU alert from xymon I’d like to see the current time interval’s CPU stats (The third line of top display - %us, %sy, %ni, etc) rather than the averages of those stats since the last reboot of the machine. This is an old ‘complaint’ in bugzilla which was closed in 2007 as ‘not a bug’. I couldn’t find any work-arounds or customizing tips out there to change the display in xymon. What changes can a person make to see the current time interval (any top display other than the first one) statistics for the cpu display in xymon? What about changes in the monitored OS? Any suggestions are appreciated! Thanks. > I think I know what you are asking for, but to be sure can you define what you mean by "current time interval" ? Thanks!

4 9

best (or any) way to remember disabled tests on the main page?
by ohemming＠gmail.com 06 Aug '14

06 Aug '14

I'm not really sure how to explain this but hopefully it will make sense Let's say the main Xymon page is divided up as follows: prod --> 8 servers devel --> 8 servers One of the 'prod' servers has a known issue causing it to go red, so we disable the test and it switches to blue. Is there any way to reflect this on the main page (which has now, of course, gone back to green)? Ideally, I'd like to see the name of the server group ("prod" in the example) change to blue from white on the main view to remind me there's an ignored test. But I don't want the "main view" colour to change from green Possible? Easy? Thanks

3 4

Problem with conn test
by S.Pelzetter＠wvs-basa.de 06 Aug '14

06 Aug '14

Hello, I have expanded the number of clients to test with xymon putting some internet-vpn-router of our company. Now I have the problem, that the conn tests very often failing on some hosts, generating a lot of alert e-mails. But if I use a terminal on the xymon server and ping the clients "manually" there is no timeout. I wonder why xymon is detecting a timeout for this hosts. The xymon server is a vmware virtual machine with ubuntu 12.04 LTS server on an esx 5.0.0 latest patches. This morning I've update xymon from 4.3.10 to 4.3.17 and updated the vmware-tools on that machine, but the problem still remains. What happend to that? Thanks. -- This email was Anti Virus checked by Astaro Security Gateway. http://www.sophos.com

2 1

monitoring virtual vips
by rsimion＠transunion.com 06 Aug '14

06 Aug '14

Greetings, I have virtual vips to monitor -- there will be no client installed on these virtual vips, obviously. I would like to monitor the ping test and see if ports are open. Here is what I have in the hobbit-clients.cfg PORT LOCAL=%(:23456|:8377|:8378|:23455) $" state=LISTEN SUBPAGE=vips I'm still getting only a dash on the vips page -- thanks in advance for any help Rod

2 2

xymon-rclient
by jlaidman＠rebel-it.com.au 06 Aug '14

06 Aug '14

Hi No, it only works through a shell running on the target device. It sounds like what you want is prolly available with the http or post content check built into Xymon. If the XML is sent via a post message then you can define a "post" tag in hosts.cfg. Check out the "HTTP Tests" section in the man page. https://www.xymon.com/help/manpages/man5/hosts.cfg.5.html. Or if you need more control over the content being sent (eg custom headers) you can define your own protocol in protocols.cfg. https://www.xymon.com/help/manpages/man5/protocols.cfg.5.html Or for maximum flexibility, create an ext script to run on the Xymon server. J On 06/08/2014 8:17 PM, "Rod Simioni" <rod.simioni at gmail.com> wrote: > Hi, > > Can I use the xymon-rclient to send xml to the remote client -- actually a > virtual IP? > > I don't need to login into the remote client, I just need to send xml via > http to the remote client and if the return data is good, then it is green, > if it is bad, it will be red. > > Can it do this? Thanks in advance. > > Rod >

1 0

external script not working
by rsimion＠transunion.com 06 Aug '14

06 Aug '14

My custom script is not reporting anything to the Hobbit Server. I'm running this script on the Hobbit server. I have placed my script in /usr/lib/hobbit/server/ext/b.pl. The script basically does nothing but report green: #!/usr/bin/perl -w use strict; # Xymon Variables #$ENV{BBPROG} = "b.pl"; my($TESTNAME) = "test"; my($BBHOME) = $ENV{BBHOME}; my($BB) = $ENV{BB}; my($BBDISP) = $ENV{BBDISP}; my($BBVAR) = $ENV{BBVAR}; my($MACHINE) = $ENV{MACHINE}; my($DATE) = localtime; my($COLOR) = "green"; my($MSG) = ""; my($HEAD) = ""; my($DATA) = ""; print "$MACHINE\n"; $MACHINE =~ s/\./,/g; my($cmd) = "$BB $BBDISP \"status $MACHINE.$TESTNAME $COLOR $DATE $HEAD\n$DATA\n$MSG\""; system($cmd); I put the following lines in the /etc/hobbit/hobbitlaunch.cfg [test] ENVFILE /usr/lib/hobbit/server/etc/hobbitserver.cfg NEEDS hobbitd CMD /usr/lib/hobbit/server/ext/b.pl LOGFILE $BBSERVERLOGS/test.log INTERVAL 5m I have put 'test' in the bb-hosts file: subpage vips VIPS xxx.xxx.xxx.xxx server1 #test xxx.xxx.xxx.xxx server2 #test Any help will be greatly appreciated. Rod

2 3

dynamic logfile name
by r.oliver＠web.de 06 Aug '14

06 Aug '14

Hello! I have an application that creates everyday a new logfile and I want to monitor it. The manpage for client-local.cfg says: The file:FILENAME line defines the filename of the file to monitor. As with the "log:" entries, a filename enclosed in backticks means a command which will generate the filenames dynamically. My question is: From what version on does xymon support this feature? I'm currently running 4.2.3. And can someone give me an example for the syntax with backticks? Thanks in advance! Oliver P.S. I hope that this time my message does not count as reply to any previous message.

3 5

Xymongrep in 4.3.17 does not load local hosts.cfg
by wk.fdmdb＠gmail.com 04 Aug '14

04 Aug '14

2014-07-29 18:53 GMT+02:00 J.C. Cleaver <cleaver at terabithia.org>: > > > On Tue, July 29, 2014 1:20 am, Wladimir K wrote: >> Hi, >> >> I compiled xymon 4.3.17 on Suse 11 distribution successfuly and the >> server is running. But when I call xymongrep I get the following: >> >> xymon at n1000 > xymongrep test >> 2014-07-29 09:24:01 Cannot load >> /opt/xymon/default/server/etc/hosts.cfg, or file is empty >> >> Xymon is installed in /opt/xymon/default/. >> >> The file exists and has the necessery permissions: >> >> -rw-rw-r-- 1 xymon users 1308 2014-07-18 14:45 >> /opt/xymon/default/server/etc/hosts.cfg >> >> The hosts.cfg is also not empty and worked with Xymon 4.2.3. >> >> With strace I don't see that xymongrep even tries to open the file: >> > *snip* >> >> Actually, instead of opening the file it is trying to get it via network: >> socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 3 fcntl(3, F_SETFL, >> O_RDONLY|O_NONBLOCK) = 0 connect(3, {sa_family=AF_INET, >> sin_port=htons(1984), sin_addr=inet_addr("127.0.0.1")}, 16) = -1 >> EINPROGRESS (Operation now in progress) >> select(4, [3], [3], NULL, {15, 0}) = 1 (out [3], left {14, 999998}) >> getsockopt(3, SOL_SOCKET, SO_ERROR, [72058139498774528], [4]) = 0 >> write(3, "config hosts.cfg", 16) = 16 >> >> Which does not work as the port is not 1984 but 1988. But why does it >> try to make network connection and doesn't open the local file. And >> even then it writes to stdout that it tries to open the file locally >> but it doesn't. >> >> Thanks in advance for your help. >> >> WK > > I'd have to validate, however it seems like xymongrep might do that when > it's running without an environment. Typically its use case is in scripts > that are being executed via xymonlaunch and have already had an "ENV > /path/to/xymonserver.cfg" (or equivalent) set up for them. This would also > explain not hitting your custom XYMONDPORT value. > > From the command line, can you try running 'xymoncmd > --env=/path/to/xymonservercfg xymongrep test', or explicitly specifying > 'env HOSTSCFG=/path/to/hosts.cfg xymongrep test'? > > > HTH, > -jc > Thanks, that helped! Best regards Wladimir

1 0

transform REMOTE_USER for display purposes?
by rlhamil2＠gmail.com 03 Aug '14

03 Aug '14

There are those who have asked for a way to transform the REMOTE_USER variable when it's used for display purposes (like in the enadis.sh CGI program). This can be perhaps more desirable when client certs are used with +FakeBasicAuth; the rather long identifying strings then used in the xymonpasswd (or comparable) file are a bit ugly. AFAIK, Apache's mod_env will not modify standard CGI environment variables; so the CGI's would have to do it. If they checked if some optional RE was in a config file, they could use that to convert REMOTE_USER into something better suited to display than e.g. /CN=CAcert WoT User/emailAddress=johndoe at nobody.com (trivial example of what a free cert from CAcert might show up as); or there could be a file that just mapped REMOTE_USER values to display names. Overkill, or worthwhile? :-)

2 1

Confusion on Alerts
by skadz＠skadz.com 01 Aug '14

01 Aug '14

Hi Folks - I'm a little confused to why something is happening in my alerts config, so I thought I would ask for some help to clear up the confusion. I have a specific set of systems that I only want to alert on if the procs alarm is on for 30 mins or more, while my default is 6 minutes (1 more than the poller). So I have 2 things (well, for this, I have a whole lot more doing other things): # only alarm after 30 min machines HOST=%echo.*.domain.com SERVICE=procs MAIL $EMAIL_OPS DURATION>30m REPEAT=60m MAIL $OPS_PAGER_1 DURATION>30m REPEAT=10m COLOR=red STOP HOST=* MAIL $EMAIL_OPS DURATION>6m REPEAT=60m MAIL $OPS_PAGER_1 DURATION>6m REPEAT=10m COLOR=red I _THOUGHT_ that if the STOP keyword was in there, it would stop processing after that, but when I look in the web interface, I see this: procs bb-notice at domain.com 30m 1s - 1h - purple,yellow,red pager at domain.com,pagerbox at domain.com (S) 30m 1s - 10m - red bb-notice at domain.com 6m 1s - 1h - purple,yellow,red pager at domain.com,pagerbox at domain.com 6m 1s - 10m - red The lines after the first set are red (which makes me think it's pointing out bad configuration?). What am I missing here? Thanks! Skadz

2 2