flushing rrd files - now possible via rrdcached
There now exists a patch to rrdtool (which has already been merged) that allows xymon to work with the following config-changes:
[rrdcache] LOGFILE $XYMONSERVERLOGS/rrdcached.log NEEDS xymond CMD $RRD_BASE/bin/rrdcached -w 1800 -z 900 -f 1800 -l $XYMONVAR/rrdcached/rrdcached.socket -j $XYMONVAR/rrdcached -p $XYMONVAR/rrdcached/rrdcached.pid -g -l 0.0.0.0:41984 -b $XYMONVAR/ -B
[rrdstatus] CMD +--no-cache [rrddata] CMD +--no-cache
And in the ENVFILE /etc/xymon/xymonserver.cfg add: RRDCACHED_ADDRESS="127.0.0.1:41984" RRDCACHED_STRIPPATH="$XYMONVAR"
these are there to override the default rrd install location
so that it uses the head build from rrdtool
RRD_BASE="/opt/rrdtool-1.4.999" LD_LIBRARY_PATH="$RRD_BASE/lib/"
The patch to rrdtool: https://github.com/oetiker/rrdtool-1.x/pull/462 has already been incorporated with the main development branch, so I hope it will get into the next rrdtool-1.4 release.
Ciao, Martin
From: Xymon [mailto:xymon-bounces at xymon.com] On Behalf Of Martin Sperl Sent: Dienstag, 29. April 2014 13:03 To: Jeremy Laidman Cc: xymon at xymon.com Subject: Re: [Xymon] flushing rrd files
Thanks - the problem is that those sockets change so you have to "guess" their path and see if it is functional (as with your perl example - in my case there are actually 16 socket files in temp and I have to iterate over all of them to find the 2 that are actually working - slightly stupid...)
Obviously this also does not work from a remote node which only has access to the rrd files via NFS... Meaning I have to write a http proxy for that...
So I was wondering if xymond_rrd with the --no-cache would use the RRDCache if set in the environment. My guess would be I need to use a different env file which has RRDCACHED_ADDRESS set correctly (which rrdtool makes use of automatically if the ENV is set) and then configure xymond_rrd to use that environment instead.... If that works then that rrd-specific caching could get removed as a whole - with the exception of the need for
I just want to avoid testing this on our live system, so I was asking if anyone has experience with this...
Martin
From: Jeremy Laidman [mailto:jlaidman at rebel-it.com.au] Sent: Dienstag, 29. April 2014 04:33 To: Martin Sperl Cc: xymon at xymon.com<mailto:xymon at xymon.com> Subject: Re: [Xymon] flushing rrd files
I think work would need to be done to integrate with rrdcached.
The RRD stats are cached by xymond_rrd. The showgraph.cgi binary sends a flush command to the xymond_rrd instances, via UNIX sockets, requesting a cache flush, so that graphs are up-to-date. You could probably emulate this by running showgraph.cgi from the command-line. Like so:
SCRIPT_NAME=showgraph.sh REQUEST_METHOD=GET QUERY_STRING="host=hostname.example.com<http://hostname.example.com>&service=conn" /usr/lib/xymon/server/bin/showgraph.cgi >/dev/null
In this case, the "conn" is not used for anything (all RRD files are flushed), but has to be a (I think) valid RRD filename (without the extension) or graph name (defined in graphs.cfg).
Also, if you can send a string directly to the two UNIX sockets, you can cause a flush. The format is simply the hostname in slashes, like "/hostname.example.com<http://hostname.example.com>" and so you can probably achieve this using modern versions netcat or socat, or other things that can write to sockets.
And just for fun, here's an implementation in Perl.
Cheers Jeremy
#!/usr/bin/perl -Tw
my $hostname=shift; die "Specify hostname\n" unless defined($hostname);
my $socketdir=$ENV{XYMONTMP}; die "XYMONTMP not defined, run from xymoncmd\n" unless $socketdir;
use Socket; socket(SOCK, AF_UNIX, SOCK_DGRAM, 0) or die "socket: $!\n"; use Fcntl qw(F_GETFL F_SETFL O_NONBLOCK); my $flags = fcntl(SOCK, F_SETFL, O_NONBLOCK) or die "fcntl: $!\n";
opendir(DIR,$socketdir) or die "$!: $socketdir\n"; while(my $socketfile=readdir(DIR)) { next unless substr($socketfile,0,7) eq "rrdctl."; my $socketpath="$socketdir/$socketfile"; if (! -e $socketpath) { warn "not found: $socketpath\n"; next; } if (! -w $socketpath) { warn "not writeable: $socketpath\n"; next; } if (! -S $socketpath) { warn "not a socket: $socketpath\n"; next; }
$socketaddr=sockaddr_un($socketpath);
if (defined(send(SOCK, "/$hostname/", 0, $socketaddr))) {
print "Flush command for '$hostname' sent to $socketfile\n";
} else {
warn "Flush command for '$hostname' failed to send to socket $socketfile\n";
}
}
On 25 April 2014 22:29, Martin Sperl <Martin.Sperl at amdocs.com<mailto:Martin.Sperl at amdocs.com>> wrote: Hi!
Is there a means to flush the rrd files filled in from xymond_rrd say via the "xymon" command?
As an alternative is it possible to use the generic RRDCACHED as a replacement for the caching? Any experience with that?
Thanks, Martin
This message and the information contained herein is proprietary and confidential and subject to the Amdocs policy statement, you may review at http://www.amdocs.com/email_disclaimer.asp
Xymon mailing list Xymon at xymon.com<mailto:Xymon at xymon.com> http://lists.xymon.com/mailman/listinfo/xymon
Hello,
I have more and more targets to handle through one xymonproxy and seems to make the main xymon server crash since a week..
I have tried to lower xymonserver.cfg values ( MAXMSG_*) but without any success.
I'm using XYMON 4.3.12 on Solaris 10.2.
The logs say on the main server:
calling ftok('/project/xymon0/refer/xymon_cur-vers/server',4) ftok() returns: 0x400FD56 Could not get shm of size 2621440: No such file or directory
Semaphore wait aborted: Invalid argument Semaphore wait aborted: Identifier removed
How Solaris 10 handles ipcs: https://www.princeton.edu/~unix/Solaris/troubleshoot/ipc.html#project
IPCS Parameters for xymond (prctl)
project.max-shm-memory privileged 6.04GB - deny - system 16.0EB max deny - project.max-shm-ids privileged 128 - deny - system 16.8M max deny - project.max-sem-ids privileged 128 - deny - system 16.8M max deny - process.max-sem-nsems privileged 512 - deny - system 32.8K max deny - process.max-sem-ops privileged 512 - deny - system 2.15G max deny - -
ipcs -a
Message Queues:
T ID KEY MODE OWNER GROUP CREATOR CGROUP
NATTCH SEGSZ CPID LPID ATIME DTIME CTIME
Shared Memory:
m 16777303 0x900fd56 --rw------- 6adm 6adm 6adm 6adm
1 131072 22594 18524 8:52:48 11:23:06 8:52:48
m 16777302 0x800fd56 --rw------- 6adm 6adm 6adm 6adm
2 614400 22594 18524 8:52:48 11:23:06 8:52:48
m 16777301 0x700fd56 --rw------- 6adm 6adm 6adm 6adm
3 2048000 22594 18524 8:52:48 11:23:06 8:52:48
m 16777300 0x600fd56 --rw------- 6adm 6adm 6adm 6adm
1 614400 22594 18524 8:52:48 11:23:06 8:52:48
m 16777299 0x500fd56 --rw------- 6adm 6adm 6adm 6adm
1 614400 22594 18524 8:52:48 11:23:06 8:52:48
m 16777298 0x400fd56 --rw------- 6adm 6adm 6adm 6adm
6 2048000 22594 18524 8:52:48 11:23:06 8:52:48
m 16777297 0x300fd56 --rw------- 6adm 6adm 6adm 6adm
2 614400 22594 18524 8:52:48 11:23:06 8:52:48
m 16777296 0x200fd56 --rw------- 6adm 6adm 6adm 6adm
2 2048000 22594 18524 8:52:48 11:23:06 8:52:48
m 16777295 0x100fd56 --rw------- 6adm 6adm 6adm 6adm
3 2048000 22594 18524 8:52:48 11:23:06 8:52:48
T ID KEY MODE OWNER GROUP CREATOR CGROUP
NSEMS OTIME CTIME
Semaphores:
s 16777303 0x900fd56 --ra------- 6adm 6adm 6adm 6adm
3 no-entry 8:52:48
s 16777302 0x800fd56 --ra------- 6adm 6adm 6adm 6adm
3 11:25:56 8:52:48
s 16777301 0x700fd56 --ra------- 6adm 6adm 6adm 6adm
3 11:26:44 8:52:48
s 16777300 0x600fd56 --ra------- 6adm 6adm 6adm 6adm
3 no-entry 8:52:48
s 16777299 0x500fd56 --ra------- 6adm 6adm 6adm 6adm
3 no-entry 8:52:48
s 16777298 0x400fd56 --ra------- 6adm 6adm 6adm 6adm
3 11:26:45 8:52:48
s 16777297 0x300fd56 --ra------- 6adm 6adm 6adm 6adm
3 11:26:44 8:52:48
s 16777296 0x200fd56 --ra------- 6adm 6adm 6adm 6adm
3 11:26:16 8:52:48
s 16777295 0x100fd56 --ra------- 6adm 6adm 6adm 6adm
3 11:26:45 8:52:48
Cordialement, Regards,Mit freundlichen Grüßen,
Gautier BEGIN
On 7 May 2014 19:36, Gautier Begin <gbegin at csc.com> wrote:
calling ftok('/project/xymon0/refer/xymon_cur-vers/server',4) ftok() returns: 0x400FD56 *Could not get shm of size 2621440: No such file or directory*
Does this exist: /project/xymon0/refer/xymon_cur-vers/
Is it readable by the Xymon user?
Cheers Jeremy
Hello,
Yes it is is. This is the root directory of the xymon server.
The server crashed again this night. I have got this message in the xymonlaunch log 2014-05-08 00:10:14 Fatal error in select: Invalid argument 2014-05-08 00:10:14 Cannot open checkpoint file /project/xymon0/refer/xymon_cur-vers/server/tmp/xymond.chk.1399500614 : Too many open files
Then all channels log write: 2014-05-08 00:10:14 Tried to down BOARDBUSY: Invalid argument 8941 2014-05-08 00:10:14 Semaphore wait aborted: Invalid argument 8941 2014-05-08 00:10:14 Semaphore wait aborted: Invalid argument
Current limit on open files is 4096 . But currently, xymonlaunch is using only 3 and xymond 4. The other xymon channel processes are using 4. So no more than 100.
Cordialement, Regards,Mit freundlichen Grüßen,
Gautier BEGIN
System Tools Team Lead CACEIS and APERAM accounts CSC Computer Sciences Luxembourg S.A. 12D Impasse Drosbach L-1882 Luxembourg
Global Outsourcing Service | p:+352 24 834 276 | m:+352 621 229 172 | gbegin at csc.com | www.csc.com
CSC • This is a PRIVATE message. If you are not the intended recipient, please delete without copying and kindly advise us by e-mail of the mistake in delivery. NOTE: Regardless of content, this e-mail shall not operate to bind CSC to any order or other contract unless pursuant to explicit written agreement or government initiative expressly permitting the use of e-mail for such purpose • CSC Computer Sciences SAS • Registered Office: Immeuble Le Balzac, 10 Place des Vosges, 92072 Paris La Défense Cedex, France • Registered in France: RCS Nanterre B 315 268 664
From: Jeremy Laidman <jlaidman at rebel-it.com.au> To: Gautier Begin/LUX/CSC at CSC Cc: "xymon at xymon.com" <xymon at xymon.com> Date: 05/08/2014 02:22 AM Subject: Re: [Xymon] XYMOND Crash with IPCS issue
On 7 May 2014 19:36, Gautier Begin <gbegin at csc.com> wrote: calling ftok('/project/xymon0/refer/xymon_cur-vers/server',4) ftok() returns: 0x400FD56 Could not get shm of size 2621440: No such file or directory
Does this exist: /project/xymon0/refer/xymon_cur-vers/
Is it readable by the Xymon user?
Cheers Jeremy
A piece of infromation more:
I have to empty the server/tmp directory where checkpoints are to be able to restart the xymon.
Cordialement, Regards,Mit freundlichen Grüßen,
Gautier BEGIN
System Tools Team Lead CACEIS and APERAM accounts CSC Computer Sciences Luxembourg S.A. 12D Impasse Drosbach L-1882 Luxembourg
Global Outsourcing Service | p:+352 24 834 276 | m:+352 621 229 172 | gbegin at csc.com | www.csc.com
CSC • This is a PRIVATE message. If you are not the intended recipient, please delete without copying and kindly advise us by e-mail of the mistake in delivery. NOTE: Regardless of content, this e-mail shall not operate to bind CSC to any order or other contract unless pursuant to explicit written agreement or government initiative expressly permitting the use of e-mail for such purpose • CSC Computer Sciences SAS • Registered Office: Immeuble Le Balzac, 10 Place des Vosges, 92072 Paris La Défense Cedex, France • Registered in France: RCS Nanterre B 315 268 664
From: Gautier Begin/LUX/CSC at CSC To: Jeremy Laidman <jlaidman at rebel-it.com.au> Cc: "xymon at xymon.com" <xymon at xymon.com> Date: 05/08/2014 08:52 AM Subject: Re: [Xymon] XYMOND Crash with IPCS issue Sent by: "Xymon" <xymon-bounces at xymon.com>
Hello,
Yes it is is. This is the root directory of the xymon server.
The server crashed again this night. I have got this message in the xymonlaunch log 2014-05-08 00:10:14 Fatal error in select: Invalid argument 2014-05-08 00:10:14 Cannot open checkpoint file /project/xymon0/refer/xymon_cur-vers/server/tmp/xymond.chk.1399500614 : Too many open files
Then all channels log write:
2014-05-08 00:10:14 Tried to down BOARDBUSY: Invalid argument
8941 2014-05-08 00:10:14 Semaphore wait aborted: Invalid argument
8941 2014-05-08 00:10:14 Semaphore wait aborted: Invalid argument
Current limit on open files is 4096 . But currently, xymonlaunch is using only 3 and xymond 4. The other xymon channel processes are using 4. So no more than 100.
Cordialement, Regards,Mit freundlichen Grüßen,
Gautier BEGIN
System Tools Team Lead CACEIS and APERAM accounts CSC Computer Sciences Luxembourg S.A. 12D Impasse Drosbach L-1882 Luxembourg
Global Outsourcing Service | p:+352 24 834 276 | m:+352 621 229 172 | gbegin at csc.com | www.csc.com
CSC • This is a PRIVATE message. If you are not the intended recipient, please delete without copying and kindly advise us by e-mail of the mistake in delivery. NOTE: Regardless of content, this e-mail shall not operate to bind CSC to any order or other contract unless pursuant to explicit written agreement or government initiative expressly permitting the use of e-mail for such purpose • CSC Computer Sciences SAS • Registered Office: Immeuble Le Balzac, 10 Place des Vosges, 92072 Paris La Défense Cedex, France • Registered in France: RCS Nanterre B 315 268 664
From: Jeremy Laidman <jlaidman at rebel-it.com.au> To: Gautier Begin/LUX/CSC at CSC Cc: "xymon at xymon.com" <xymon at xymon.com> Date: 05/08/2014 02:22 AM Subject: Re: [Xymon] XYMOND Crash with IPCS issue
On 7 May 2014 19:36, Gautier Begin <gbegin at csc.com> wrote: calling ftok('/project/xymon0/refer/xymon_cur-vers/server',4) ftok() returns: 0x400FD56 Could not get shm of size 2621440: No such file or directory
Does this exist: /project/xymon0/refer/xymon_cur-vers/
Is it readable by the Xymon user?
Cheers Jeremy
Xymon mailing list Xymon at xymon.com http://lists.xymon.com/mailman/listinfo/xymon
Hello,
I finally solved the problem. XYMON crash was due to a data over flow coming from the proxy. So the solution was to regulate data flow on the proxy by tuning MAXMSGSPERCOMBO and SLEEPBETWEENMSGS in the xymonserver.cfg file. Unfortunately, the man page of xymonserver.cfg is much less explicit than in the xymonnet one on how to use these values.
So I lowered the MAXMSGSPERCOMBO and raised the SLEEPBETWEENMSGS and that solved the problem.
MAXMSGSPERCOMBO="50" # Default 100 - 0 =>unlimited SLEEPBETWEENMSGS="5000" # microseconds
The result can be seen in the graph of the xymonproxy test display and the graph of the xymond test display.
Remain that during the high flow period, the xymonnet on the proxy doesn't send any data. Should I continue to lower MAXMSGSPERCOMBO and raise the SLEEPBETWEENMSGS ?
Cordialement, Regards,Mit freundlichen Grüßen,
Gautier BEGIN
System Tools Team Lead CACEIS and APERAM accounts CSC Computer Sciences Luxembourg S.A. 12D Impasse Drosbach L-1882 Luxembourg
Global Outsourcing Service | p:+352 24 834 276 | m:+352 621 229 172 | gbegin at csc.com | www.csc.com
CSC • This is a PRIVATE message. If you are not the intended recipient, please delete without copying and kindly advise us by e-mail of the mistake in delivery. NOTE: Regardless of content, this e-mail shall not operate to bind CSC to any order or other contract unless pursuant to explicit written agreement or government initiative expressly permitting the use of e-mail for such purpose • CSC Computer Sciences SAS • Registered Office: Immeuble Le Balzac, 10 Place des Vosges, 92072 Paris La Défense Cedex, France • Registered in France: RCS Nanterre B 315 268 664
From: Gautier Begin/LUX/CSC at CSC To: "xymon at xymon.com" <xymon at xymon.com> Cc: "Xymon" <xymon-bounces at xymon.com> Date: 05/08/2014 10:22 AM Subject: Re: [Xymon] XYMOND Crash with IPCS issue Sent by: "Xymon" <xymon-bounces at xymon.com>
A piece of infromation more:
I have to empty the server/tmp directory where checkpoints are to be able to restart the xymon.
Cordialement, Regards,Mit freundlichen Grüßen,
Gautier BEGIN
System Tools Team Lead CACEIS and APERAM accounts CSC Computer Sciences Luxembourg S.A. 12D Impasse Drosbach L-1882 Luxembourg
Global Outsourcing Service | p:+352 24 834 276 | m:+352 621 229 172 | gbegin at csc.com | www.csc.com
CSC • This is a PRIVATE message. If you are not the intended recipient, please delete without copying and kindly advise us by e-mail of the mistake in delivery. NOTE: Regardless of content, this e-mail shall not operate to bind CSC to any order or other contract unless pursuant to explicit written agreement or government initiative expressly permitting the use of e-mail for such purpose • CSC Computer Sciences SAS • Registered Office: Immeuble Le Balzac, 10 Place des Vosges, 92072 Paris La Défense Cedex, France • Registered in France: RCS Nanterre B 315 268 664
From: Gautier Begin/LUX/CSC at CSC To: Jeremy Laidman <jlaidman at rebel-it.com.au> Cc: "xymon at xymon.com" <xymon at xymon.com> Date: 05/08/2014 08:52 AM Subject: Re: [Xymon] XYMOND Crash with IPCS issue Sent by: "Xymon" <xymon-bounces at xymon.com>
Hello,
Yes it is is. This is the root directory of the xymon server.
The server crashed again this night. I have got this message in the xymonlaunch log 2014-05-08 00:10:14 Fatal error in select: Invalid argument 2014-05-08 00:10:14 Cannot open checkpoint file /project/xymon0/refer/xymon_cur-vers/server/tmp/xymond.chk.1399500614 : Too many open files
Then all channels log write:
2014-05-08 00:10:14 Tried to down BOARDBUSY: Invalid argument
8941 2014-05-08 00:10:14 Semaphore wait aborted: Invalid argument
8941 2014-05-08 00:10:14 Semaphore wait aborted: Invalid argument
Current limit on open files is 4096 . But currently, xymonlaunch is using only 3 and xymond 4. The other xymon channel processes are using 4. So no more than 100.
Cordialement, Regards,Mit freundlichen Grüßen,
Gautier BEGIN
System Tools Team Lead CACEIS and APERAM accounts CSC Computer Sciences Luxembourg S.A. 12D Impasse Drosbach L-1882 Luxembourg
Global Outsourcing Service | p:+352 24 834 276 | m:+352 621 229 172 | gbegin at csc.com | www.csc.com
CSC • This is a PRIVATE message. If you are not the intended recipient, please delete without copying and kindly advise us by e-mail of the mistake in delivery. NOTE: Regardless of content, this e-mail shall not operate to bind CSC to any order or other contract unless pursuant to explicit written agreement or government initiative expressly permitting the use of e-mail for such purpose • CSC Computer Sciences SAS • Registered Office: Immeuble Le Balzac, 10 Place des Vosges, 92072 Paris La Défense Cedex, France • Registered in France: RCS Nanterre B 315 268 664
From: Jeremy Laidman <jlaidman at rebel-it.com.au> To: Gautier Begin/LUX/CSC at CSC Cc: "xymon at xymon.com" <xymon at xymon.com> Date: 05/08/2014 02:22 AM Subject: Re: [Xymon] XYMOND Crash with IPCS issue
On 7 May 2014 19:36, Gautier Begin <gbegin at csc.com> wrote: calling ftok('/project/xymon0/refer/xymon_cur-vers/server',4) ftok() returns: 0x400FD56 Could not get shm of size 2621440: No such file or directory
Does this exist: /project/xymon0/refer/xymon_cur-vers/
Is it readable by the Xymon user?
Cheers Jeremy
Xymon mailing list Xymon at xymon.com http://lists.xymon.com/mailman/listinfo/xymon
Xymon mailing list Xymon at xymon.com http://lists.xymon.com/mailman/listinfo/xymon
On 7 May 2014 16:54, Martin Sperl <Martin.Sperl at amdocs.com> wrote:
There now exists a patch to rrdtool (which has already been merged) that allows xymon to work with the following config-changes:
This is very cool. What were the changes? Was it to support the RRDCACHED_* variables?
J
No – 2 different places needed a change:
· Template support when using rrdcached (essentially translating templates to full updates)
· RRDCACHED_STRIPPATH support to strip the “leading” path making it a relative path which is supported by rrdcached (besides when using sockets)
RRDCACHED_ADDRESS already existed as a transparent option.
These obviously could also get patched with xymon, but as this also is beneficial for other tools (cacti,…) I decided to patch rrdtool.
One note though: rrdtool in the master branch comes with a new feature called “skip-past-updates” (rrdrool update -- skip-past-updates …), this remains unsupported when using rrdcached.
Ciao, Martin
From: Jeremy Laidman [mailto:jlaidman at rebel-it.com.au] Sent: Donnerstag, 08. Mai 2014 01:55 To: Martin Sperl Cc: xymon at xymon.com Subject: Re: [Xymon] flushing rrd files - now possible via rrdcached
On 7 May 2014 16:54, Martin Sperl <Martin.Sperl at amdocs.com<mailto:Martin.Sperl at amdocs.com>> wrote: There now exists a patch to rrdtool (which has already been merged) that allows xymon to work with the following config-changes:
This is very cool. What were the changes? Was it to support the RRDCACHED_* variables?
J
This message and the information contained herein is proprietary and confidential and subject to the Amdocs policy statement, you may review at http://www.amdocs.com/email_disclaimer.asp
participants (3)
-
gbegin@csc.com
-
jlaidman@rebel-it.com.au
-
Martin.Sperl@amdocs.com