July 2006 - Xymon - Xymon mailman web

[hobbit] TCP/IP stats (bits/s) limited to 100M
by henrik＠hswn.dk 10 Jul '06

10 Jul '06

On Wed, Jun 28, 2006 at 04:07:44PM +0200, Nicolas Dorfsman wrote: > >The RRD files are created as "DERIVE" datatypes with a minimum > >value of > >0, which should handle 32/64-bit counter overflows automatically. > >(See the rrdcreate man-page). > > Well...the man is not so confident : > > If you cannot tolerate ever mistaking the > occasional counter reset for a legitimate > counter wrap, and would prefer "Unknowns" > for all legitimate counter wraps and resets, > always use DERIVE with min=0. Otherwise, > using COUNTER with a suitable max will > return correct values for all legitimate > counter wraps, mark some counter resets as > "Unknown", but can mistake some counter > resets for a legitimate counter wrap. OK, you got me on that one. It seems that using COUNTER for the byte-counts in both the netstat- and ifstat-RRD's might be a good idea. The question then becomes "what's a suitable max" for these data ? Should I assume they are 32-bit counters ? I know some of them are not (e.g. Solaris has 64-bit counters for bytes in/out per interface). I'll change it to a counter now, with MAX set to "unknown". The overflow handling should still work correctly, if I understand the RRD docs right. Note: This doesn't affect all of the existing RRD's, only new ones created. Regards, Henrik

3 3

[hobbit] [patch] bb-findhost.cgi redirects to the page when only one host matches
by henrik＠hswn.dk 09 Jul '06

09 Jul '06

On Wed, Jun 28, 2006 at 04:03:11PM +0200, Charles Goyard wrote: > I made another modification, so that when all matches are on the same > page, you get to that page at once. [snip] > There's two flavours of the patch made against today's snapshot (Chances > are you will find a better way to implement it : > > - you get to the page without the host anchor ; > - you get to the first match. I chose the second option, implemented it slightly differently, and made an option on the findhost form so you can choose whether or not to jump directly to the host or go via the choose-page. Regards, Henrik

1 0

[hobbit] ghost clients reporting
by henrik＠hswn.dk 09 Jul '06

09 Jul '06

On Thu, Jun 29, 2006 at 02:39:05PM +0200, Stephane Caminade wrote: > I've just added the BBGHOST=2 setting to my hobbitserver.cfg file > (running Hobbit Monitor 4.2-beta-20060605), and I was wondering if it is > possible to create a warning in the hobbitd test when a ghost is detected ? Not right now. The ghost reports do not change the color of the hobbitd test, as you found out. I could make that an option, but I'll leave that for later. > By the way, what does the BBGHOST=0 option do ? > BBGHOSTS="0" # How to handle status messages from unknown hosts. > # 0=accept message, 1=discard message, 2=discard message and log the event > > If I use the setting 0, does the message get added somewhere (and if so, > where ...) ? It's a left-over from the older Big Brother days, and it doesn't really make much sense. In fact, I've removed the BBGHOSTS setting from the code recently and made it always behave as BBGHOSTS=2. There's also a specific "Ghost report" tool now so you can view the ghosts sorted by hostname, IP or age. Regards, Henrik

1 0

[hobbit] Problem with the latest release
by henrik＠hswn.dk 09 Jul '06

09 Jul '06

On Thu, Jun 29, 2006 at 11:22:33PM -0400, Wes Neal wrote: > I had 4.1.1 running fine and wanted to move to the latest beta release. The > compile seemed to go fine, but the hobbit log shows this now: > > 2006-06-29 22:16:33 Could not get shm of size 262144: Invalid argument > 2006-06-29 22:16:33 Cannot setup status channel My Solaris 8 man-page says: EINVAL The size argument is less than the system-imposed minimum or greater than the system-imposed maximum. Cannot be this, since your shm settings are for a size between 1 and 16000000 - the Hobbit request of 262144 bytes fits this. EINVAL A shared memory identifier exists for key but the size of the segment associated with it is less than size and size is not equal to 0. Sounds like this is the one. Your old Hobbit may still be running ? Login as the Hobbit user and do an "ipcs -m". If there are any segments listed, either Hobbit is running or they weren't cleaned up properly upon shutdown. In that case, you can remove them with "ipcrm". Henrik

1 0

[hobbit] testing pulldata in hobbit
by henrik＠hswn.dk 09 Jul '06

09 Jul '06

On Thu, Jun 29, 2006 at 08:12:16PM -0500, Scheblein, Adam wrote: > I have been testing the new pulldata option in the snapshot from June > 29, 2006, however I have run into a problem. Basically, hobbit has > decided to not pull data from 2 of my 5 clients that I am testing this > on. This is in a trucluster 5 environment and its cluster mate is also > having the same problems. The "pulldata" code had not been tested much except in a very limited lab-setup. Over the past couple of days I have done some more extensive testing with help from the Hobbit user who requested this feature, and the current snapshots should work a lot better. Regards, Henrik

1 0

[hobbit] bbcmd on clients does not set MACHINE and MACHINEDOTS
by frederic.mangeant＠steria.com 09 Jul '06

09 Jul '06

Dominique Frise a écrit : > Hi, > > We noticed a problem when using bbcmd on clients for testing extension > scripts: > the variables exported in runclient.sh are not set. > > At least MACHINE and MACHINEDOTS should be set for beeing able to test > any client-side extension script (Big Brother style or Hobbit style). Hi Henrik I totally agree with Dominique : I'm currently "upgrading" Big Brother clients to Hobbit 4.1.2p1 clients, and some of my external scripts are run from other scripts (post-backup ones, for example). Would it be possible to modify bbcmd to export MACHINE, MACHINEDOTS, BBOSTYPE and HOBBITCLIENTHOME, so we can run external scripts outside of Hobbit ? Thanks in advance ! -- Frédéric Mangeant Steria EDC Sophia-Antipolis

3 2

Hobbit alerts with qmail failing
by vadud3＠gmail.com 09 Jul '06

09 Jul '06

Hi All Anyone running qmail on hobbit server? I am not receiving alerts but web page get updated fine. Here are the related info on my setup MAIL="sendmail -f user at host.com" HOST=%jump(.*)|rusty(.*) MAIL pager at host.com 1.2.0.195 rusty.host.com # ssh echo this is a test email | sendmail -f user at host.com (works). /var/log/qmail/current shows the email too. The following command does not generate an email bb 127.0.0.1 "status rusty.home.com.conn red `date` this is a test alert" I see the web page got updated with red alert. But /var/log/qmail/current does not show any email generated either. Is the "MAIL" variable defined wrong then? -- Asif Iqbal PGP Key: 0xE62693C5 KeyServer: pgp.mit.edu

2 2

iostat graphs?
by Dominique.Frise＠unil.ch 07 Jul '06

07 Jul '06

Hi, This question is for Henrik ;-) SunOS and Irix clients collect iostat data but this is not yet exploited on the server. When are the iostat graphs to be expected and how about other OS support like Linux and Darwin? Thanks, Dominique UNIL - University of Lausanne

1 0

hobbitd_alert --test color matching
by Dominique.Frise＠unil.ch 07 Jul '06

07 Jul '06

Hi, We cannot test our rules with other color than red: the matching messages always show same results for the "red" color: [bb at iris bin]$ ./hobbitd_alert --test bilbo msgs red | egrep "Match with|Mail alert|Script alert" 00017141 2006-07-06 11:05:04 *** Match with 'PAGE=%^$ *** 00017141 2006-07-06 11:05:04 *** Match with '$MAIL_DH' *** 00017141 2006-07-06 11:05:04 Mail alert with command 'mail -s "Hobbit [12345] bilbo:msgs CRITICAL (RED)" xxx.yyy at sample.com' 00017141 2006-07-06 11:05:04 *** Match with 'HOST=$ADMIN_HOSTS' *** 00017141 2006-07-06 11:05:04 *** Match with '$SMS_DH' *** 00017141 2006-07-06 11:05:04 Script alert with command '/soft/pub/BB/hobbit/server/bin/sendsms' and recipient 0123456789 00017141 2006-07-06 11:05:04 *** Match with 'HOST=*' *** 00017141 2006-07-06 11:05:04 *** Match with '$MAIL_ROOT UNMATCHED' *** [bb at iris bin]$ ./hobbitd_alert --test bilbo msgs yellow | egrep "Match with|Mail alert|Script alert" 00017239 2006-07-06 11:05:56 *** Match with 'PAGE=%^$ *** 00017239 2006-07-06 11:05:56 *** Match with '$MAIL_DH' *** 00017239 2006-07-06 11:05:56 Mail alert with command 'mail -s "Hobbit [12345] bilbo:msgs CRITICAL (RED)" xxx.yyy at sample.com' 00017239 2006-07-06 11:05:56 *** Match with 'HOST=$ADMIN_HOSTS' *** 00017239 2006-07-06 11:05:56 *** Match with '$SMS_DH' *** 00017239 2006-07-06 11:05:56 Script alert with command '/soft/pub/BB/hobbit/server/bin/sendsms' and recipient 0123456789 00017239 2006-07-06 11:05:56 *** Match with 'HOST=*' *** 00017239 2006-07-06 11:05:56 *** Match with '$MAIL_ROOT UNMATCHED' *** Same results with "purple" color. PS: this is 4.2-beta-10060601 Any ideas? Dominique UNIL - University of Lausanne

1 1

Confreport Question
by allan＠zandahar.net 06 Jul '06

06 Jul '06

Hi all, Just been looking at the confreport for some hosts and noticed that it doesnt show that there are any port or proc tests and also shows the default disk and cpu limits instead of the ones set for the particular host. I know these other values are correct and working as one of them has a massive drive array thats on 97% but still has a huge chunk of space free and its still green as the limit is 98 and 99. Question is basically...should the custom value be shown here or not ? If theyre not then maybe they can be ? <--insert groveling to Henrik here If they should display...then plz help!! :) Thanks Allan

2 3