bb2 problem with pageset ?
list Beau Olivier
Hi, i'm running RC, and have noticed on the bb2 page of my UNIX pageset, that i have history event of all, not just of my unix servers... Is this a bug ? Olivier
list Henrik Størner
▸
On Thu, Jul 20, 2006 at 12:18:57PM +0200, Beau Olivier wrote:
i'm running RC, and have noticed on the bb2 page of my UNIX pageset, that i have history event of all, not just of my unix servers... Is this a bug ?
No, the bb2 page shows all hosts defined in Hobbit that are non-green, and the eventlog does the same. Regards, Henrik
list Beau Olivier
Has that changed ? it wasnt like this before (4.1.2p2) let me re-explain, i have a global bb2.html and unix pageset and a nt pageset that i use for bb2.html only. on my unix pageset bb2.html page, i have alerts of unix servers and eventlog of unix AND nt servers i pretty confused.. olivier
▸
-------- Message d'origine--------
De: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid]
Date: jeu. 20/07/2006 12:41
À: user-ae9b8668bcde@xymon.invalid
Objet : Re: [hobbit] bb2 problem with pageset ?
On Thu, Jul 20, 2006 at 12:18:57PM +0200, Beau Olivier wrote:i'm running RC, and have noticed on the bb2 page of my UNIX pageset, that i have history event of all, not just of my unix servers... Is this a bug ?
No, the bb2 page shows all hosts defined in Hobbit that are non-green, and the eventlog does the same. Regards, Henrik
list Francesco Duranti
Same here now that I'm checking. The bb2.html page for pageset is correct but the events in the bottom is reporting all events and not only those of the pageset. Francesco -----Messaggio originale----- Da: Beau Olivier [mailto:user-9358a2fdac7a@xymon.invalid] Inviato: giovedì 20 luglio 2006 12.54 A: user-ae9b8668bcde@xymon.invalid Oggetto: [hobbit] RE : [hobbit] bb2 problem with pageset ?
▸
Has that changed ? it wasnt like this before (4.1.2p2)
let me re-explain,
i have a global bb2.html and unix pageset and a nt pageset that i use for bb2.html only.
on my unix pageset bb2.html page, i have alerts of unix servers and eventlog of unix AND nt servers
i pretty confused..
olivier
-------- Message d'origine--------
De: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid]
Date: jeu. 20/07/2006 12:41
À: user-ae9b8668bcde@xymon.invalid
Objet : Re: [hobbit] bb2 problem with pageset ?
On Thu, Jul 20, 2006 at 12:18:57PM +0200, Beau Olivier wrote:i'm running RC, and have noticed on the bb2 page of my UNIX pageset, that i have history event of all, not just of my unix servers... Is this a bug ?
No, the bb2 page shows all hosts defined in Hobbit that are non-green, and the eventlog does the same. Regards, Henrik
list Nicolas Lienard
Hi,
On the server side, in hobbit-clients.cfg, i ve this:
DEFAULT
UP 1h
LOAD 5.0 10.0
DISK * 90 95
MEMPHYS 100 101
MEMSWAP 60 80
MEMACT 95 97
PROC %(/usr/sbin/cron|crond).* 1 5
The problem is that some hosts have more than 5 cron running.
So i declared before DEFAULT section, a specific host to override the
global PROC values:
For instance, the host is called "mrtg-wat":
HOST=mrtg-wat
PROC %(/usr/sbin/cron|crond).* 1 20
But it doesn't work, Hobbit takes the 2 values as 2 differents check.
%(/usr/sbin/cron|crond).* (found 26, req. between 1 and 20)
%(/usr/sbin/cron|crond).* (found 26, req. between 1 and 5)
Is there a way to change this behaviour in order to override the DEFAULT
value for PROC ?
Hobbit version: hobbit-4.1.2p1
Thanks
Nicolas
list Henrik Størner
▸
On Thu, Jul 20, 2006 at 04:02:14PM +0200, Nicolas wrote:
DEFAULT
UP 1h
LOAD 5.0 10.0
DISK * 90 95
MEMPHYS 100 101
MEMSWAP 60 80
MEMACT 95 97
PROC %(/usr/sbin/cron|crond).* 1 5Putting PROC entries in the DEFAULT section is generally a bad idea, unless you never - ever - want to have exceptions from the rule. I'd suggest removing it from the default section, and using this instead: HOST=* PROC %(/usr/sbin/cron|crond).* 1 5 EXHOST=mrtg-wat PROC %(/usr/sbin/cron|crond).* 1 HOST=mrtg-wat or some similar configuration.
Hobbit version: hobbit-4.1.2p1
Ouch, there's some problems with the config file parsing in the 4.1.2 releases. I'm not entirely sure the above will work in 4.1.2p1; you may have to list the hosts explicitly for this. It's fixed in the 4.2 release candidate currently being tested. Regards, Henrik
list Henrik Størner
▸
On Thu, Jul 20, 2006 at 12:54:03PM +0200, Beau Olivier wrote:
Has that changed ? it wasnt like this before (4.1.2p2) let me re-explain, i have a global bb2.html and unix pageset and a nt pageset that i use for bb2.html only. on my unix pageset bb2.html page, i have alerts of unix servers and eventlog of unix AND nt servers
OK, I understand what you mean now - and yes, it seems this has changed
as the eventlog code was generalized into being used also for a stand-
alone reporting tool.
The attached patch should re-instate the old behaviour, please let me
know if it works.
Regards,
Henrik
-------------- next part --------------
--- lib/eventlog.c 2006/06/01 12:29:16 1.35
+++ lib/eventlog.c 2006/07/20 21:51:41
@@ -98,7 +98,8 @@
char *pageregex, char *expageregex,
char *hostregex, char *exhostregex,
char *testregex, char *extestregex,
- char *colrregex, int ignoredialups)
+ char *colrregex, int ignoredialups,
+ f_hostcheck hostcheck)
{
FILE *eventlog;
char eventlogfilename[PATH_MAX];
@@ -220,6 +221,7 @@
(eventhost && !bbh_item(eventhost, BBH_FLAG_NOBB2)) &&
(wanted_eventcolumn(svcname)) ) {
if (ignoredialups && bbh_item(eventhost, BBH_FLAG_DIALUP)) continue;
+ if (hostcheck && (hostcheck(hostname) == NULL)) continue;
if (pageregexp) {
char *pagename = bbh_item(eventhost, BBH_PAGEPATH);
--- lib/eventlog.h 2006/06/01 12:29:36 1.9
+++ lib/eventlog.h 2006/07/20 21:51:49
@@ -24,6 +24,7 @@
struct event_t *next;
} event_t;
+typedef void * (*f_hostcheck)(char *hostname);
extern char *eventignorecolumns;
extern int havedoneeventlog;
@@ -32,6 +33,7 @@
char *pagematch, char *expagematch,
char *hostmatch, char *exhostmatch,
char *testmatch, char *extestmatch,
- char *colormatch, int ignoredialups);
+ char *colormatch, int ignoredialups,
+ f_hostcheck hostcheck);
#endif
--- web/bb-eventlog.c 2006/07/11 17:18:22 1.34
+++ web/bb-eventlog.c 2006/07/20 21:46:42
@@ -132,7 +132,7 @@
fprintf(stdout, "<center>\n");
do_eventlog(stdout, maxcount, maxminutes, fromtime, totime,
pageregex, expageregex, hostregex, exhostregex, testregex, extestregex,
- colrregex, ignoredialups);
+ colrregex, ignoredialups, NULL);
fprintf(stdout, "</center>\n");
headfoot(stdout, "event", "", "footer", COL_GREEN);
--- bbdisplay/pagegen.c 2006/07/20 16:06:41 1.176
+++ bbdisplay/pagegen.c 2006/07/20 21:41:50
@@ -986,7 +986,7 @@
/* Dont redo the eventlog or acklog things */
if (strcmp(p, "eventlog.sh") == 0) {
if (bb2eventlog && !havedoneeventlog) do_eventlog(output, bb2eventlogmaxcount, bb2eventlogmaxtime,
- NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, bb2nodialups);
+ NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, bb2nodialups, find_host);
}
else if (strcmp(p, "acklog.sh") == 0) {
if (bb2acklog && !havedoneacklog) do_acklog(output, bb2acklogmaxcount, bb2acklogmaxtime);
@@ -1163,7 +1163,7 @@
do_bb2ext(output, "BBMKBB2EXT", "mkbb");
/* Dont redo the eventlog or acklog things */
- if (bb2eventlog && !havedoneeventlog) do_eventlog(output, bb2eventlogmaxcount, bb2eventlogmaxtime, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, bb2nodialups);
+ if (bb2eventlog && !havedoneeventlog) do_eventlog(output, bb2eventlogmaxcount, bb2eventlogmaxtime, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, bb2nodialups, find_host);
if (bb2acklog && !havedoneacklog) do_acklog(output, bb2acklogmaxcount, bb2acklogmaxtime);
}
list Nicolas Lienard
HI,
Thanks for answear.
Indeed, it doesn't seem to work correctly with this hobbit version:
HOST=*
PROC %(/usr/sbin/cron|crond).* 1 5 EXHOST=mrtg-wat,test-host
PROC %(/usr/sbin/cron|crond).* 1 30 HOST=mrtg-wat
PROC %(/usr/sbin/cron|crond).* 1 6 HOST=test-host
The results for test-host on PROC was:
&unknown %(/usr/sbin/cron|crond).* (found 6, req. between 1 and 5)
green %(/usr/sbin/cron|crond).* (found 6, req. between 1 and 30)
green %(/usr/sbin/cron|crond).* (found 6, req. between 1 and 6)
Other servers which are'nt in EXHOST have this behaviour:
green %(/usr/sbin/cron|crond).* (found 1, req. between 1 and 5)
green %(/usr/sbin/cron|crond).* (found 1, req. between 1 and 30)
green %(/usr/sbin/cron|crond).* (found 1, req. between 1 and 6)
As you suggested, i changed conf like this (list the hosts explicitly) :
HOST=*
PROC %(/usr/sbin/cron|crond).* 1 5 EXHOST=mrtg-wat,gie-s2if-front2
HOST=test-host
PROC %(/usr/sbin/cron|crond).* 1 6
HOST=mrtg-wat
PROC %(/usr/sbin/cron|crond).* 1 30
it seems to work but there is a strange behaviour for the global rule and
EXHOST:
test-host PROC show this:
&unknown %(/usr/sbin/cron|crond).* (found 6, req. between 1 and 5)
green %(/usr/sbin/cron|crond).* (found 6, req. between 1 and 6)
Others hosts are OK (only the global PROC).
Do you know what the "unknown" state means ? is it obstructing ?
Thanks
Nicolas
▸
Le Jeu 20 juillet 2006 23:06, Henrik Stoerner a écrit :On Thu, Jul 20, 2006 at 04:02:14PM +0200, Nicolas wrote:DEFAULT UP 1h LOAD 5.0 10.0 DISK * 90 95 MEMPHYS 100 101 MEMSWAP 60 80 MEMACT 95 97 PROC %(/usr/sbin/cron|crond).* 1 5Putting PROC entries in the DEFAULT section is generally a bad idea, unless you never - ever - want to have exceptions from the rule. I'd suggest removing it from the default section, and using this instead: HOST=* PROC %(/usr/sbin/cron|crond).* 1 5 EXHOST=mrtg-wat PROC %(/usr/sbin/cron|crond).* 1 HOST=mrtg-wat or some similar configuration.Hobbit version: hobbit-4.1.2p1Ouch, there's some problems with the config file parsing in the 4.1.2 releases. I'm not entirely sure the above will work in 4.1.2p1; you may have to list the hosts explicitly for this. It's fixed in the 4.2 release candidate currently being tested. Regards, Henrik
list Beau Olivier
works !! thanks :) olivier -----Message d'origine----- De : Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid] Envoyé : jeudi 20 juillet 2006 23:57 À : user-ae9b8668bcde@xymon.invalid Objet : Re: [hobbit] RE : [hobbit] bb2 problem with pageset ? On Thu, Jul 20, 2006 at 12:54:03PM +0200, Beau Olivier wrote:
Has that changed ? it wasnt like this before (4.1.2p2) let me re-explain, i have a global bb2.html and unix pageset and a nt pageset that i use for bb2.html only. on my unix pageset bb2.html page, i have alerts of unix servers and eventlog of unix AND nt servers
OK, I understand what you mean now - and yes, it seems this has changed as the eventlog code was generalized into being used also for a stand- alone reporting tool. The attached patch should re-instate the old behaviour, please let me know if it works. Regards, Henrik
list Beau Olivier
Hi, i'ld like to have custom graph using NCVq with a GAUGE, but hobbit creates the rrd with DERIVE.. here my definition: NCV_sessions="current_sessions:GAUGE,maxlimit:NONE" is this a bug ? olivier
list Charles Goyard
▸
Beau Olivier a écrit :
i'ld like to have custom graph using NCVq with a GAUGE, but hobbit creates the rrd with DERIVE.. here my definition: NCV_sessions="current_sessions:GAUGE,maxlimit:NONE"
That's because hobbit strips under_scores from the DS name. Try : NCV_sessions="currentsessions:GAUGE,maxlimit:NONE" -- Charles Goyard - user-98f9625a7a59@xymon.invalid - (+33) 1 45 38 01 31
list Beau Olivier
works !! thanks :) olivier -----Message d'origine----- De : Charles Goyard [mailto:user-98f9625a7a59@xymon.invalid] Envoyé : jeudi 27 juillet 2006 10:22 À : user-ae9b8668bcde@xymon.invalid Objet : Re: [hobbit] graph ncv problem with RC ? Beau Olivier a écrit :
i'ld like to have custom graph using NCVq with a GAUGE, but hobbit creates the rrd with DERIVE.. here my definition: NCV_sessions="current_sessions:GAUGE,maxlimit:NONE"
That's because hobbit strips under_scores from the DS name. Try : NCV_sessions="currentsessions:GAUGE,maxlimit:NONE" -- Charles Goyard - user-98f9625a7a59@xymon.invalid - (+33) 1 45 38 01 31