badTEST:x:y:z not working
list Michael Heinecke
Dear all, looks to me, as if the "badTEST:x:y:z" tags are not working. I.e. "badssh:1:2:3" turns immediatley red if ssh goes down at the monitored host. I've looked into bbtest-net.c and figured out, that .. - At line 775, the for loop << "for (swalk=svchead; (swalk && strcmp(swalk->testname, testname)); swalk = swalk->next) >> compares swalk->testname and testname ---> which are never equal, because the string testname always includes the :x.y:z settings and swalk->testname (of course) not. Hmmm... your experiences? I've found no posting about this kind of issue on the list, and i think "badTEST" is one of the most important tags. I'am using 4.1.1 (4.1.2 (snapshot 09.08.05 also testet - same effect) @ Linux (Debian sarge and Solaris 8 :-) Thx! Michael Michael Heinecke CAX / UNIX&VoD HanseNet Telekommunikation GmbH Überseering 33 a, 22297 Hamburg Telefon: +49 (0)40 23726-2768 Telefax: +49 (0)40 23726-3485 http://www.alice-dsl.de <http://www.alice-dsl.de/>; , http://www.hansenet.de <http://www.hansenet.de/>;
list Henrik Størner
▸
On Thu, Aug 11, 2005 at 01:46:16PM +0200, user-c9ab24a501e1@xymon.invalid wrote:
Dear all, looks to me, as if the "badTEST:x:y:z" tags are not working. I.e. "badssh:1:2:3" turns immediatley red if ssh goes down at the monitored host.
There has been some reports about this, but I never got around to looking into exactly what caused it.
▸
I've looked into bbtest-net.c and figured out, that .. - At line 775, the for loop << "for (swalk=svchead; (swalk && strcmp(swalk->testname, testname)); swalk = swalk->next) >> compares swalk->testname and testname ---> which are never equal, because the string testname always includes the :x.y:z settings and swalk->testname (of course) not.
Aha - that would explain it.
I think this patch should fix it, then.
Henrik
-------------- next part --------------
--- bbnet/bbtest-net.c 2005/08/10 15:13:21 1.220
+++ bbnet/bbtest-net.c 2005/08/11 13:42:56
@@ -773,9 +773,12 @@
if (timespec) inscope = periodcoversnow(timespec);
if (strlen(testname) && badcounts && inscope) {
+ char *p;
twalk = NULL;
+ p = strchr(testname, ':'); if (p) *p = '\0';
for (swalk=svchead; (swalk && strcmp(swalk->testname, testname)); swalk = swalk->next) ;
+ if (p) *p = ':';
if (swalk) {
if (swalk == httptest) twalk = h->firsthttp;
else if (swalk == ldaptest) twalk = h->firstldap;
list Frédéric Mangeant
▸
Henrik Stoerner a écrit :
On Thu, Aug 11, 2005 at 01:46:16PM +0200, user-c9ab24a501e1@xymon.invalid wrote:Dear all, looks to me, as if the "badTEST:x:y:z" tags are not working. I.e. "badssh:1:2:3" turns immediatley red if ssh goes down at the monitored host.There has been some reports about this, but I never got around to looking into exactly what caused it.
Hi Henrik I have 2 questions about badTEST : - if I want a test to go yellow at the 1st failure and red at the 2nd, is the syntax "badTEST:1:1:2" correct ? I had problems with bbgen 2.16 (some tests went red at the 1st failure). With the patch you posted and latest 4.1.2 snapshot, I get this yellow for 00:53 and then red Is it because "bbretest" runs every minute ? - if I want ALL my network tests to go yellow at the 1st failure and red at the 2nd, can I add this at the top of my bb-hosts : 0.0.0.0 .default. # badconn:1:1:2 badftp:1:1:2 badtelnet:1:1:2 [...] Thanks in advance for your answers. -- Frédéric Mangeant Steria EDC Sophia-Antipolis
list Eric van de Meerakker
Hello Henrik, Just to confirm this: your patch seems to be working OK. I set "badhttp:0:1:2" and have now seen the http test go from green to yellow before turning red. Thanks! This will keep down the number of false alerts. Eric mail to : user-004bb4e22963@xymon.invalid phone : +XX (X)XX XXX XXXX fax : +XX (X)XX XXX XXXX
▸
-----Original Message-----
From: Henrik Stoerner [mailto:user-ce4a2c883f75@xymon.invalid]
Sent: Thursday, August 11, 2005 3:49 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: Re: [hobbit] badTEST:x:y:z not working
On Thu, Aug 11, 2005 at 01:46:16PM +0200, user-c9ab24a501e1@xymon.invalid wrote:Dear all, looks to me, as if the "badTEST:x:y:z" tags are not working. I.e. "badssh:1:2:3" turns immediatley red if ssh goes down at the monitored host.
There has been some reports about this, but I never got around to looking into exactly what caused it.
I've looked into bbtest-net.c and figured out, that .. - At line 775, the for loop << "for (swalk=svchead; (swalk && strcmp(swalk->testname, testname)); swalk = swalk->next) >> compares swalk->testname and testname ---> which are never equal, because the string testname always includes the :x.y:z settings and swalk->testname
(of course) not. Aha - that would explain it. I think this patch should fix it, then. Henrik