From eponymousalias at yahoo.com Thu Jan 1 01:28:22 2015 From: eponymousalias at yahoo.com (eponymous alias) Date: Wed, 31 Dec 2014 16:28:22 -0800 Subject: check_icmp fails to compile on Solaris 10 and 11 (#1317) In-Reply-To: Message-ID: <1420072102.58631.YahooMailBasic@web160702.mail.bf1.yahoo.com> MSG_CONFIRM seems to be a Linux-ism, and not a historical one at that. The sendmsg(2) man page says: MSG_CONFIRM (Linux 2.3+ only) Tell the link layer that forward progress happened: you got a suc- cessful reply from the other side. If the link layer doesn't get this it will regularly reprobe the neighbour (e.g. via a unicast ARP). Only valid on SOCK_DGRAM and SOCK_RAW sockets and currently only implemented for IPv4 and IPv6. See arp(7) for details. So I wouldn't blame Oracle for this. (There is no MSG_CONFIRM symbol defined in the Solaris /usr/include/sys/socket.h header.) The OpenGroup year-2013 page on sendmsg() doesn't mention it, either: http://pubs.opengroup.org/onlinepubs/9699919799/functions/sendmsg.html The construction around this line in plugins-root/check_icmp.c: 883 len = sendmsg(sock, &hdr, MSG_CONFIRM); will need to change to make the code portable to other platforms. -------------------------------------------- On Wed, 12/31/14, doombirdAD wrote: Subject: check_icmp fails to compile on Solaris 10 and 11 (#1317) To: "Monitoring Plugins Development" Date: Wednesday, December 31, 2014, 11:52 AM We're trying to update our monitoring in our mostly Solaris 10 and 11 environment, but I started getting errors compiling check_icmp in the most recent version. I suspect it's an issue between glibc and Oracle's libc. From Solaris 10: gcc -DLOCALEDIR=\"/usr/local/share/locale\" -DHAVE_CONFIG_H -I. -I.. -I.. -I../lib -I../gl -I../intl -I../plugins -I/include -D_REENTRANT -I/include -DNP_VERSION='"2.1.1"' -g -O2 -MT check_icmp.o -MD -MP -MF .deps/check_icmp.Tpo -c -o check_icmp.o check_icmp.c check_icmp.c: In function `send_icmp_ping': check_icmp.c:883: error: `MSG_CONFIRM' undeclared (first use in this function) check_icmp.c:883: error: (Each undeclared identifier is reported only once check_icmp.c:883: error: for each function it appears in.) check_icmp.c: In function `recvfrom_wto': check_icmp.c:943: error: structure has no member named `msg_control' check_icmp.c:944: error: structure has no member named `msg_controllen' *** Error code 1 make: Fatal error: Command failed for target `check_icmp.o' Also errors when compiling from Solaris 11: gcc -DLOCALEDIR=\"/usr/local/share/locale\" -DHAVE_CONFIG_H -I. -I.. -I.. -I../lib -I../gl -I../intl -I../plugins -I/usr/include -D_REENTRANT -DNP_VERSION='"2.1.1"' -g -O2 -MT check_icmp.o -MD -MP -MF .deps/check_icmp.Tpo -c -o check_icmp.o check_icmp.c check_icmp.c: In function ?send_icmp_ping?: check_icmp.c:883:28: error: ?MSG_CONFIRM? undeclared (first use in this function) check_icmp.c:883:28: note: each undeclared identifier is reported only once for each function it appears in check_icmp.c: In function ?recvfrom_wto?: check_icmp.c:943:5: error: ?struct msghdr? has no member named ?msg_control? check_icmp.c:944:5: error: ?struct msghdr? has no member named ?msg_controllen? check_icmp.c:948:11: warning: assignment makes pointer from integer without a cast check_icmp.c:948:45: warning: assignment makes pointer from integer without a cast check_icmp.c:952:4: warning: passing argument 2 of ?memcpy? makes pointer from integer without a cast /usr/include/iso/string_iso.h:54:14: note: expected ?const void *? but argument is of type ?int? *** Error code 1 make: Fatal error: Command failed for target `check_icmp.o' ? Reply to this email on GitHub. From notifications at github.com Fri Jan 2 20:23:35 2015 From: notifications at github.com (doombirdAD) Date: Fri, 02 Jan 2015 11:23:35 -0800 Subject: check_icmp fails to compile on Solaris 10 and 11 (#1317) In-Reply-To: References: Message-ID: I found the fix in 2 hacks: 1. `MSG_CONFIRM` is not defined in socket.h on Solaris. I added an `#ifndef` line to define it at the beginning of check_icmp.c. There appears to be a section devoted to "sometimes undefined system macros" so that seems to be the right place for it. 2. `_XOPEN_SOURCE` needs to be defined to use the version of the msghdr struct that has "msg_control" and "msg_controllen". I don't know where the "standardized" place to put this would be. I just added a `check_icmp_CFLAGS = $(AM_CFLAGS) -D_XOPEN_SOURCE=500` to the plugins-root/Makefile.am. I didn't know what else might be affected by adding that definition, so I limited scope to just the file that needs it. With these two fixes, I was able to compile on both 10 and 11. I also compiled on my CentOS 7 box and was able to build RPMs and run check_icmp without issue. -- Reply to this email on GitHub: https://github.com/monitoring-plugins/monitoring-plugins/issues/1317#issuecomment-68553876 -------------- next part -------------- An HTML attachment was scrubbed... URL: From notifications at github.com Fri Jan 16 23:15:43 2015 From: notifications at github.com (Peter Bieringer) Date: Fri, 16 Jan 2015 14:15:43 -0800 Subject: check_dns: add AAAA,MX,TXT,SRV,CNAME (#1185) In-Reply-To: References: Message-ID: take care when takeover, looks like the upstream version has a strange upcase issue: /usr/lib64/nagios/plugins/check_dns -q SOA google.de DNS OK: 0,023 seconds response time. GOOGle.de returns ns2.google.com|time=0,022598s;;;0,000000 /usr/lib64/nagios/plugins/check_dns -q MX redhat.com DNS OK: 0,043 seconds response time. REDHAt.com returns 10 mx2.REDHAt.com.,5 mx1.REDHAt.com.|time=0,042732s;;;0,000000 (amout of upcase chars vary....haven't digged through the code how this could happen...perhaps one can inform upstream) # /usr/lib64/nagios/plugins/check_dns -V check_dns v2.0.1 (nagios-plugins 2.0.1) -- Reply to this email on GitHub: https://github.com/monitoring-plugins/monitoring-plugins/pull/1185#issuecomment-70331446 -------------- next part -------------- An HTML attachment was scrubbed... URL: From waja at cyconet.org Sat Jan 17 10:20:01 2015 From: waja at cyconet.org (Jan Wagner) Date: Sat, 17 Jan 2015 10:20:01 +0100 Subject: check_dns: add AAAA,MX,TXT,SRV,CNAME (#1185) In-Reply-To: References: Message-ID: <54BA2941.4040307@cyconet.org> Hi Peter, thanks for bringing this up to our attention. > take care when takeover, looks like the upstream version has a strange > upcase issue: nagios-plugins isn't our upstream, but they forked us (not digging into details here again). So we can share code in some areas in each direction. > /usr/lib64/nagios/plugins/check_dns -q SOA google.de > DNS OK: 0,023 seconds response time. GOOGle.de returns > ns2.google.com|time=0,022598s;;;0,000000 > > /usr/lib64/nagios/plugins/check_dns -q MX redhat.com > DNS OK: 0,043 seconds response time. REDHAt.com returns 10 > mx2.REDHAt.com.,5 mx1.REDHAt.com.|time=0,042732s;;;0,000000 > > (amout of upcase chars vary....haven't digged through the code how this > could happen...perhaps one can inform upstream) We have spotted some quirks with this changes too, which is one reason that we don't have pulled it (yet). > /usr/lib64/nagios/plugins/check_dns -V > > check_dns v2.0.1 (nagios-plugins 2.0.1) Maybe you might want to report this issue at https://github.com/nagios-plugins/nagios-plugins/issues/new With kind regards, Jan. From notifications at github.com Tue Jan 20 14:25:41 2015 From: notifications at github.com (jyoung15) Date: Tue, 20 Jan 2015 05:25:41 -0800 Subject: check_snmp default timeout incorrect in --help output (#1318) Message-ID: running check_snmp -h shows the following -t, --timeout=INTEGER Seconds before connection times out (default: 10) In check_snmp.c, this comes from DEFAULT_SOCKET_TIMEOUT: printf (UT_CONN_TIMEOUT, DEFAULT_SOCKET_TIMEOUT); However the actual timeout is defined in DEFAULT_TIMEOUT (which is 1 second, not 10) timeout_interval = DEFAULT_TIMEOUT; -- Reply to this email on GitHub: https://github.com/monitoring-plugins/monitoring-plugins/issues/1318 -------------- next part -------------- An HTML attachment was scrubbed... URL: From notifications at github.com Tue Jan 20 14:27:37 2015 From: notifications at github.com (jyoung15) Date: Tue, 20 Jan 2015 05:27:37 -0800 Subject: check_icmp fails to compile on Solaris 10 and 11 (#1317) In-Reply-To: References: Message-ID: Same issue affects FreeBSD. MSG_CONFIRM is not defined. -- Reply to this email on GitHub: https://github.com/monitoring-plugins/monitoring-plugins/issues/1317#issuecomment-70652578 -------------- next part -------------- An HTML attachment was scrubbed... URL: From notifications at github.com Wed Jan 21 12:37:30 2015 From: notifications at github.com (Lars Vogdt) Date: Wed, 21 Jan 2015 03:37:30 -0800 Subject: Rewrite of check_ircd.pl (#1319) Message-ID: I did a rewrite of the check_ircd.pl script to allow the script to * allow connections using SSL * provide performance data The options are downwards compatible and the script itself is used in our infrastructure since more than a year now without any problems. An example of a command definition: ```bash define command { command_name check_ircd_ssl command_line $USER1$/check_ircd -H $HOSTADDRESS$ -w $ARG1$ -c $ARG2$ -s -p 6697 } ``` Please note that instead of the perl(Socket) module, the new script needs perl(IO::Socket::INET6). Left TODO: IPv6 support. You can view, comment on, or merge this pull request online at: https://github.com/monitoring-plugins/monitoring-plugins/pull/1319 -- Commit Summary -- * - rewrite of the plugin to support connections via SSL * enhance header section * - whitespace fix only * - prepare for release -- File Changes -- M plugins-scripts/check_ircd.pl (359) -- Patch Links -- https://github.com/monitoring-plugins/monitoring-plugins/pull/1319.patch https://github.com/monitoring-plugins/monitoring-plugins/pull/1319.diff -- Reply to this email on GitHub: https://github.com/monitoring-plugins/monitoring-plugins/pull/1319 -------------- next part -------------- An HTML attachment was scrubbed... URL: From notifications at github.com Sat Jan 24 17:48:14 2015 From: notifications at github.com (squareatom) Date: Sat, 24 Jan 2015 08:48:14 -0800 Subject: Don't let check_disk hang on hanging file systems (6986aa1) In-Reply-To: References: Message-ID: check_disk still hangs indefinitely when there is a problem with network mounts such as nfs. For example, if you were to run df -t nfs and it hangs, this plugin would also hang. This is at least true for Linux (RHWS 3,4, RHEL 5,6) but I imagine for other *nix'es as well. -- Reply to this email on GitHub: https://github.com/monitoring-plugins/monitoring-plugins/commit/6986aa1d0a352d8d02eed4896034631fffd25a27#commitcomment-9428666 -------------- next part -------------- An HTML attachment was scrubbed... URL: From s.gangale at twt.it Mon Jan 26 16:57:55 2015 From: s.gangale at twt.it (Sara Gangale) Date: Mon, 26 Jan 2015 15:57:55 -0000 Subject: Help - check_http plugin Message-ID: <00d877897d.0001d03980.0006875748.00000012db@twt.it> Good Morning, I?m currently implementing a Centreon server to monitor all our infrastructure (virtual and physical servers both windows and linux, host esx, network equipment and so on). I?m trying to use the check_http plugin to monitor a web page and I?d like the plugin to behave as follows: - Opens an URL - Sends username and password when asked - On the returned page check for content (a string contained in the HTTP answer) Everything works fine if I don?t follow redirects, so I?m able to grep a string in the first HTTPS response using the following command (sensitive data are hidden for obvious reasons): [root at centreon plugins]# ./check_http -H ?hostname? -u ?uri? -s "string1" -f follow HTTP OK: HTTP/1.1 200 OK - 38209 bytes in 0.022 second response time |time=0.021580s;;;0.000000 size=38209B;;;0 If I authenticate and set ?f follow in order to follow the redirect, something goes wrong: I?m able to grep strings in the login page, and not in what is shown AFTER the authentication: [root at centreon plugins]# ./check_http -H ?hostname? -u ?uri? -s "string2" -f follow -a username:password HTTP OK: HTTP/1.1 200 OK - 38209 bytes in 0.030 second response time |time=0.029502s;;;0.000000 size=38209B;;;0 [root at centreon plugins]# ./check_http -H ?hostname? -u ?uri? -s "string2" -f follow -a username:password HTTP CRITICAL: HTTP/1.1 200 OK - string ?string2? not found on 'http://hostname:port/uri' - 38209 bytes in 0.028 second response time |time=0.027837s;;;0.000000 size=38209B;;;0 Of course string1 is something that you can find in login page, string 2 is something you can find after the login. I don?t know if I?m not authenticating (but username and password are working and there?s no LDAP involved) or if I?m not following the redirect, because in the error I can see that the uri is always the same (HTTP CRITICAL: HTTP/1.1 200 OK - string ?string2? not found on 'http://hostname:port/uri'). How can I proceed? Is there a way to debug this? Thanks Sara Gangale System Analyst TWT S.p.A. Viale Jenner, 33 - 20159 Milano - Italy Tel +39-02-89089.288 - Fax +39-02-89089.211 Mobile +39-3371158367 E-mail: s.gangale at twt.it - Web: http://www.twt.it -------------- next part -------------- An HTML attachment was scrubbed... URL: