August 2006 - Xymon - Xymon mailman web

Disk alerts
by billieperez＠gmail.com 15 Aug '06

15 Aug '06

Hello - I'm using the big brother windows client with the hobbit server and was wondering if there was a way to specify separate mailing lists for different windows disks? Example: windows systems group gets alerted for C: and D:, application owners get alerted for E:, F: or whatever drive their application software resides on. Currently the alerts are set up where windows group gets alerted for every disk, even if it's something they dont' manage. Thanks in advance for any info.

2 2

Compile Error
by dunne＠ntms1.accuweather.com 15 Aug '06

15 Aug '06

Greetings, I am attempting to move our production Hobbit server to the latest snapshot. The latest snapshot fails to compile. The error that it gives is: gmake[1]: Entering directory `/Users/hobbit/Desktop/snapshot/bbdisplay' gcc -g -O -Wall -Wno-unused -D_REENTRANT -DBIND_8_COMPAT=1 -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -DDarwin -I. -I/Users/hobbit/Desktop/snapshot/include -I/usr/local/include -c -o bbgen.o bbgen.c In file included from bbgen.c:26: bbgen.h:173: error: conflicting types for `host_t' /usr/include/mach/mach_types.h:103: error: previous declaration of `host_t' gmake[1]: *** [bbgen.o] Error 1 gmake[1]: Leaving directory `/Users/hobbit/Desktop/snapshot/bbdisplay' gmake: *** [bbdisplay-build] Error 2 This is on an Xserve running OS X.3.9. Any advice would be greatly appreciated. Thanks, Mike Michael Dunne Systems Support Specialist AccuWeather, Inc.

1 0

Apache stats
by steve.aiello＠ge.com 15 Aug '06

15 Aug '06

I really like the graphs I get with having hobbit pull data from the mod_status module. but I was curious if there were plans to have this show up as an actual hobbit test. Would be interested in setting up thresholds on the stats provided by mod_status. I can do this via a script, but seems most of the functionality is already in hobbit, so figured I would ask. Thank you, Steve

1 0

Problems with hobbit-enadis.cgi from info page
by fduranti＠q8.it 15 Aug '06

15 Aug '06

I've recompiled a clean 4.2.0 with all the patch installed and the hobbit-enadis script is working well when called from the menu (so i think i did something wrong during the yesterday patch). I continue to get errors while enabling/disabling tests directly from the info page The test will be disabled/enabled rightly but the script will put me on this page http://itromlx10/hobbit-seccgi/(null) With an error in the browser error: Not Found The requested URL /hobbit-seccgi/(null) was not found on this server. I tested also with 4.2.0 without any patch installed and i get the same result... > Works really good ... > > Also the hobbitfetch patch is working well... > > I'm having a little problem with the enable-disable page > Called inside the info page i get a web page > http://itromlx10/hobbit-seccgi/(null) > > And the error > Not Found > The requested URL /hobbit-seccgi/(null) was not found on this server. > > > -------------------------------------------------------------- > ---------- > -------- > > Apache/2.0.55 (Red Hat) Server at itromlx10 Port 80 > > If called from a page to disable multiple time i get and > internal server error

2 1

[hobbit] Semi-unrelated question on network monitoring
by fduranti＠q8.it 15 Aug '06

15 Aug '06

I'm using Hobbit to monitor the availability of the routers and at this moment I'm using a standard mrtg setup to get data. To show data got with mrtg I'm currently using routers2.cgi, a simple perl scripts that will let you manage the visualization of data on the page (you can group routers by site or some other mechanism). You can find it at http://www.steveshipway.org/software/ it's simple to setup and really fast. The setup is simple, I'd wrote some scripts to do some thing in automatic, mostly: 1) One script that run during the night and via cfgmaker (standard mrtg script) create one mrtg.cfg file for each device i want to monitor and create history archive. 2) one script that run from hobbit every 5 minutes, get the list of all the cfg file I'm monitoring and launch mrtg with that cfg file (I need cfg files because in the past i got some problems with mrtg exiting when one error was found) 3) A script that simply send a green status to hobbit for each monitored hosts and inside that status message i've put a simple javascript that open a popup page that point directly to the routers2.cgi to open that page.... The nice thing of routers2.cgi is that he can save an archive of all your rrd file mantaining for example the 7 rrd (so you can check up to 7 day of full data), the last 4 rrd of the 1st of the month data (so you can check and compare data of each month). You can also archive single graphs, create a pdf report, do a trend statistical report to see how the usage of network is going in the next few months based on istorical data or you can directly export the data of a routers to csv file and open it with your spreadsheet to analyze what happened. It's snmpv2 compliant so no problems with gigabit interface and with cfgmaker you can automatically create the configuration file you need. There's no automatic setup so you've to create some scripting yourself but the results are good and most of all you've some really well made summary page where you can see all the interface of a router or you can create some custom graphs with the data you need ... Francesco > -----Original Message----- > From: Kauffman, Tom [mailto:KauffmanT at nibco.com] > Sent: Tuesday, August 15, 2006 3:36 PM > To: hobbit at hswn.dk > Subject: [hobbit] Semi-unrelated question on network monitoring > > > We have 14 remote sites (manufacturing facilities and distribution > warehouses) with an average of 5 Cisco switches each. We've > been using cricket to monitor the traffic on these switches, > but cricket is doing a number on the cpus of the system I'm > running it on. And I've never figured out how to get cricket > and hobbit to integrate for alerts. > > I just dropped devmon on my test system, and I really like > the integration -- but it looks like it's doing a number on > my test system cpu, and we do use the traffic graphs out of > cricket to see trends. > > So -- is cacti any lighter in it's footprint? Or easier to > keep up-to-date than cricket? Does it integrate well with hobbit? > > Should I be looking at something else? > > My current hobbit system is an ancient (as technology goes) > Dell 4300 4-way P4 - 450 system with 1 GB of memory; my test > system is a slightly newer Dell 2300 2-way P4 - 500 with 512 > MB of memory. > > Suggestions, please. > > TIA > > Tom Kauffman > NIBCO, Inc > > > CONFIDENTIALITY NOTICE: This email and any attachments are > for the exclusive and confidential use of the intended > recipient. If you are not the intended recipient, please do > not read, distribute or take action in reliance upon this > message. If you have received this in error, please notify us > immediately by return email and promptly delete this message > and its attachments from your computer system. We do not > waive attorney-client or work product privilege by the > transmission of this message. > > > > To unsubscribe from the hobbit list, send an e-mail to > hobbit-unsubscribe at hswn.dk > > >

1 0

Semi-unrelated question on network monitoring
by KauffmanT＠nibco.com 15 Aug '06

15 Aug '06

We have 14 remote sites (manufacturing facilities and distribution warehouses) with an average of 5 Cisco switches each. We've been using cricket to monitor the traffic on these switches, but cricket is doing a number on the cpus of the system I'm running it on. And I've never figured out how to get cricket and hobbit to integrate for alerts. I just dropped devmon on my test system, and I really like the integration -- but it looks like it's doing a number on my test system cpu, and we do use the traffic graphs out of cricket to see trends. So -- is cacti any lighter in it's footprint? Or easier to keep up-to-date than cricket? Does it integrate well with hobbit? Should I be looking at something else? My current hobbit system is an ancient (as technology goes) Dell 4300 4-way P4 - 450 system with 1 GB of memory; my test system is a slightly newer Dell 2300 2-way P4 - 500 with 512 MB of memory. Suggestions, please. TIA Tom Kauffman NIBCO, Inc CONFIDENTIALITY NOTICE: This email and any attachments are for the exclusive and confidential use of the intended recipient. If you are not the intended recipient, please do not read, distribute or take action in reliance upon this message. If you have received this in error, please notify us immediately by return email and promptly delete this message and its attachments from your computer system. We do not waive attorney-client or work product privilege by the transmission of this message.

1 0

Solaris Client Script and lofs File Systems
by Andy＠zespri.com 15 Aug '06

15 Aug '06

Henrik, I should have spotted this aaaaages ago, but have only just noticed while reviewing the status pages post my 4.2.0 upgrade. We use the automount facility in Solaris for home directories. So the users home is set to /home/user, but is actually contained in /export/home/user. This allows the /export/home to be shared both locally and via NFS for a central home directory across multiple servers. The automountd process automagically creates a loopback or NFS mount of /export/home/user on /home/user when the file system is accessed depending on how /etc/auto_home is configured. Which leads me to the following suggestion: as each /home/user entry shows in the mount table, but is actually the exact same information as the /export/home file system, the hobbitclient-sunos.sh script should add ^lofs to the list of file system types to exclude. I apologise if this is a bit disjointed - it's quite not as "summery" down here in the southern hemisphere as it is up there, and my head is all stuffy from a cold. Cheers, Andy. PS - apologies for the following :-( ##################################################################################### This email is intended for the person to whom it is addressed only. If you are not the intended recipient, do not read, copy or use the contents in any way. The opinions expressed may not necessarily reflect those of ZESPRI Group of Companies ('ZESPRI'). While every effort has been made to verify the information contained herein, ZESPRI does not make any representations as to the accuracy of the information or to the performance of any data, information or the products mentioned herein. ZESPRI will not accept liability for any losses, damage or consequence, however, resulting directly or indirectly from the use of this e-mail/attachments. #####################################################################################

2 1

[hobbit] Host Notes
by matthew.epp＠us.army.mil 14 Aug '06

14 Aug '06

> 6) another thing regarding notes, it should be nice if the notes system > will allow the user to put also some kind of web links on the notes > pages of hosts so that one can put links to real documentation of that > host or generic documentation regarding for example network > infrastructure into the network objects monitored). It should be also > nice if the documentation itself is saved in some kind of xml or text > file and if a html file exist in the notes directory a link to it is > visible on the notes pages... I don't know how clear this one is :P but > just to explain I'm starting to deploy cfg2html on every machine because > we need a full documentation of what's on every machine but it would be > really nice to have Since there seems to be a real need for some easily configured notes system, I decided to go ahead and create a patch file for the mod that I made. I call it "Comments" but feel free to implement it however you wish. You only need to patch one source file and recompile. This will put your notes at the bottom of every bb-hostsvc.sh page. Some of you may not want this, but I wrote this for Big Brother when there wasn't any formal notes system. If someone would like to adjust this for appearing in a different area, please do. Add these lines to the etc/hobbitserver.cfg file: BBCOMMENTS="$BBHOME/comments" # The location of the textfiles for service comments BBCOMMCONF="$BBCOMMENTS/comments.conf" # The name of the Comments configuration file Create the $BBHOME/comments directory, and then a comments.conf file in there. Here's a sample starter file: ---[SNIP]--- # This is the configuration file for the Special Instructions section on # all alert pages. To create a new alert message, put whatever # text you want into a new file, call it anything you like, but try to # stick to the convention of CI.service. Wildcards also work, but place # them at the bottom of the file, so that absolute matches can override # the wildcard entries. # NOTE: Make sure there are no trailing spaces at the end of the line. # I'll be adding in error checking code for that asap. If someone # wants to fix this, please post the code. # Portal Application Servers rules group webserver*.disk webservers.disk database*.cpu databases.cpu # # Place all wildcard matches below here # *.http ALL.http # # Point all conn tests to the same file # *.conn ALL.conn ---[SNIP]--- Then make sure to create the comment files themselves, like "ALL.conn" and put them in the same directory as the conf. You can use any html code you'd like: ---[SNIP]--- <table border=5 bordercolor=red><caption><B><font color=white>Special Instructions:</caption><tr><td><B><font color=white> Put your comments in here. Additional HTML formatting is allowed, so feel free to insert coloring, links, whatever. </td></tr></table><br> ---[SNIP]--- Now for the patch. This is the first patch file I've ever had to make, so don't scream at me too much if I didn't do this right. :) --- hobbit-4.2.0/lib/htmllog.c 2006-08-09 20:10:16.000000000 +0000 +++ hobbit-4.2.0.modified/lib/htmllog.c 2006-08-10 21:46:44.000004000 +0000 @@ -11,7 +11,7 @@ /* */ /*-------------------------------------------------------------------------- --*/ -static char rcsid[] = "$Id: htmllog.c,v 1.51 2006/05/25 14:55:41 henrik Rel $"; +static char rcsid[] = "$Id: htmllog.c,v 1.51 2006/05/25 14:55:41 henrik Exp $"; #include <ctype.h> #include <stdlib.h> @@ -22,6 +22,7 @@ #include <fcntl.h> #include <sys/types.h> #include <sys/stat.h> +#include <fnmatch.h> #include "libbbgen.h" #include "version.h" @@ -122,6 +123,105 @@ } while (restofmsg); } +/* + * Here's our function to print out the service comment file + */ +void printcomment(char *hostname, char *service) +{ +int commentstatres; +int commentretc; +int commentpage_size; +char *basename; +char *bbcomment; +char *commentpage; +char *func_matchfile = ""; +char *matchfile; +char conffile[PATH_MAX]; +char func_matchfilename[PATH_MAX] = ""; +char hostpat[PATH_MAX] = ""; +char matchfilename[PATH_MAX] = ""; +static char *bbcomments; +static char commfile[PATH_MAX] = ""; +struct stat commentstatbuf; +const char *func_hostpat; +const char *infile; +FILE *in; +FILE *bbcomm; + + if ( commentpage != NULL ) + free(commentpage); +/* Build the pattern to search for from the hostname and service */ + sprintf(hostpat, "%s.%s", hostname, service); + func_hostpat = hostpat; + +/* Search the comments.conf file for the comment file */ + + sprintf(conffile, "%s", xgetenv("BBCOMMCONF")); + + /* + * line[0] stores the input line as read, + */ + char line[1][BUFSIZ]; + infile = conffile; + + if((in = fopen(infile, "r")) == NULL){ + printf("Error opening file: %s\n", conffile); + } else { + while(fgets(line[0], BUFSIZ, in)){ + char *line_to_use = line[0]; + func_matchfile = strtok(line_to_use, " \t"); + + if(!fnmatch(func_matchfile, func_hostpat, FNM_NOESCAPE)){ + func_matchfile = strtok(NULL, " "); + strncpy(func_matchfilename, func_matchfile, strlen(func_matchfile) -1); + break; + } + } + fclose(in); + } + matchfile = func_matchfilename; +/* End of reading comments.conf */ + + sprintf(matchfilename, "%s/%s", xgetenv("BBCOMMENTS"), matchfile); + + commentstatres=stat(matchfilename, &commentstatbuf); + if( commentstatres == -1 ) { + commentpage = NULL; + } else { + commentpage_size = commentstatbuf.st_size + 4096; + + /* Fill the buffer */ + commentpage = (char *) calloc(1,commentpage_size); + + if( commentpage != NULL ) { + *commentpage = '\0'; + + /* Open and read comment file in one call*/ + if( (bbcomm = fopen(matchfilename, "r")) != NULL ) { + commentretc = fread(commentpage,commentstatbuf.st_size,1,bbcomm); + /* Was the comment file read in completely ? */ + if ( commentretc != 1 || strlen(commentpage) != commentstatbuf.st_size ) { + free(commentpage); + commentpage = NULL; + } else { +#if DEBUG + printf("comment: %s\n", commentpage); +#endif + } + fclose(bbcomm); + } else { +#if DEBUG + printf("Error opening file %s\n", matchfilename); +#endif + } + } + } + /* END OF READING IN COMMENTS FILE */ + + if (commentpage != NULL) { + printf("<BR><BR><CENTER>%s<BR></CENTER>", commentpage); + } +} void generate_html_log(char *hostname, char *displayname, char *service, char *ip, int color, char *sender, char *flags, @@ -367,6 +467,12 @@ } fprintf(output,"</CENTER>\n"); +/* + * Here we insert our service comments from the comments.conf setup + * Hope this works! + */ + printcomment(hostname, service); + headfoot(output, tplfile, "", "footer", color); }

1 0

[hobbit] Host Notes
by fduranti＠q8.it 14 Aug '06

14 Aug '06

The -m will generate really much data but i think that 1 or 2 more "w" can be good to get all the process names ... (last update on the first client where i was checking for process names I had 30 minutes trying to understand why one of my qmail process was not running :D) It could be nice to have them as some kind of configuration that can be done at configuration time or for example with the logfetch method (like the filenames or log files to check).... Can be also nice to have to choose to do network disk check with df or only local disks... This is the difference on a machine with some javavm running of the ps [bb at ITROMAS27 bb]$ ps -Aw -o pid,ppid,user,start,state,pri,pcpu,time,pmem,rsz,vsz,cmd |wc 183 2789 20249 [bb at ITROMAS27 bb]$ ps -Awww -o pid,ppid,user,start,state,pri,pcpu,time,pmem,rsz,vsz,cmd |wc 183 2801 20981 [bb at ITROMAS27 bb]$ ps -Awwwm -o pid,ppid,user,start,state,pri,pcpu,time,pmem,rsz,vsz,cmd |wc 981 20468 289819 So i think for 700 bytes we can put 2 more "w" the "m" can be problematic .... > -----Original Message----- > From: Charles Jones [mailto:jonescr at cisco.com] > Sent: Tuesday, August 15, 2006 12:28 AM > To: hobbit at hswn.dk > Subject: Re: [hobbit] Host Notes > > > > 9) .... It's possible to have some more "w" directly on the > ps command > > in the hobbit client? Something like ps -www ... Some long java > > commands are not displaying and the check on them is a bit > difficult to do ... > > > I needed this as well, I changed the client ps command in > hobbitclient-linux.sh to "ps -Awwm". The extra w gave me the > width I needed, and the m allows me to make a "TRACK" graph > of how many java threads are spawned. > > I don't think this should be the default though, as I also > ended up having to increase my MAXMSG_STATUS and MAXMSG_DATA > lengths because of the increased output generated by ps with > those options. > > -Charles > > To unsubscribe from the hobbit list, send an e-mail to > hobbit-unsubscribe at hswn.dk > > >

1 0

[hobbit] Host Notes
by fduranti＠q8.it 14 Aug '06

14 Aug '06

Just some idea for some little feature that can go in 4.2.1 :P Well before starting the list I've to say that the 4.2.0 realease is really fantastic :D With the possibility of client update directly from the server it will take just some hour to update all client or also to change the extention used or some configuration file... Thanks for the really good work !!! 1) the client Class in the info page (also some other info that just you can get on the bb-hosts and put on the info page like if the client is a pulldata client or na normal client and things like that) 2) on the status page the "Status message received from 10.202.1.201:19" is trunkated and don't show the complete port number 3) inode checking with disk check (well it should be easy and there's already the code to do the graphs) 4) the possibility for client that are not hobbit-client at this moment to send some kind of information to put on the info page (for example the version and some thing like that, useful for script that monitor storage system or oracle database to send information regarding the version of what they're monitoring) 5) in the info page the notes appear only if there's a html file in the notes directory, it should be nice to edit notes from there if there's a script to do the editing.... 6) another thing regarding notes, it should be nice if the notes system will allow the user to put also some kind of web links on the notes pages of hosts so that one can put links to real documentation of that host or generic documentation regarding for example network infrastructure into the network objects monitored). It should be also nice if the documentation itself is saved in some kind of xml or text file and if a html file exist in the notes directory a link to it is visible on the notes pages... I don't know how clear this one is :P but just to explain I'm starting to deploy cfg2html on every machine because we need a full documentation of what's on every machine but it would be really nice to have a) a Notes pages with some host related information like what the host is dedicated to, people responsible for application on that machine, telephone number and so on b) some generic link (for example grouped with CLASS definition) c) some specific link for that host d) if a hostname.html or hostname.txt is present in the notes directory a link to it (this can be usefull for cfg2html files that can be copied directly into the notes directory from some scripts). e) make the possibility to choose from the configuration if those link can be opened directly in the same page or into a new page... 7) The possibility to put one or two web page documentation link on the info page (well if the notes is done well i think there's no need for this one :P) 8) a kind of template definition for the classes ... I think i'll end to have about 10 different client setup and about 10-15 kind of file monitoring (used mostly to check the remote nfs mount point instead changing the hobbit-client shell) different template and not everytime they'll be the same so I will mix monitoring and update and I think I'll mostly end with at least 30-40 different CLASS definition... It should be nice to create a kind of template for the class definition to include file monitoring or clientversion update... 9) .... It's possible to have some more "w" directly on the ps command in the hobbit client? Something like ps -www ... Some long java commands are not displaying and the check on them is a bit difficult to do ... 10) the possibility to add some more checks to the metric page (for example I'm getting graphs for weblogic server, oracle database and netapp server and they'll not get there (I think I'll have to put some more definition on the hobbitgraphs.cfg file ... I'll check on how to do them ...) Well ... :P i've done just some requests :D Thanks for the real good work again!! Francesco > -----Original Message----- > From: Henrik Stoerner [mailto:henrik at hswn.dk] > Sent: Monday, August 14, 2006 10:21 PM > To: hobbit at hswn.dk > Subject: Re: [hobbit] Host Notes > > On Mon, Aug 14, 2006 at 12:59:10PM -0700, Charles Jones wrote: > > Darn, it seems to be gone from the latest snapshot...I want > it :-) Did > > you abandon or is it just on the backburner now? > > Not abandoned, just postponed until the 4.2 release was out. > > This week looks busy, but I'll probably do some work on it next week. > I expect it to be in 4.2.1, which will probably happen in a > months time or so as a wrap-up for any bugs that pop up with > the 4.2.0 release. > Adding a couple of harmless goodies helps to get you all to > upgrade :-) > > > Regards, > Henrik > > > To unsubscribe from the hobbit list, send an e-mail to > hobbit-unsubscribe at hswn.dk > > >

2 1