cacti-user Mailing List for Cacti (Page 11)

The forums and doc site where down over night due to mysql database 
corruption.  Everything was back up and running around 8:30 AM EST.

Tony Roman
Cacti Developer


On 8/18/15 8:24 AM, Giles Coochey wrote:
> It appears that forums.cacti.net is down?
>
> Anyone confirm or know?
>
> ------------------------------------------------------------------------------
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
>

Hi, nope, works ok,

http://downforeveryoneorjustme.com/

2015-08-18 14:24 GMT+02:00 Giles Coochey <gi...@co...>:

> It appears that forums.cacti.net is down?
>
> Anyone confirm or know?
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
>

It appears that forums.cacti.net is down?

Anyone confirm or know?

Hi All,

I submitted a Private security-related bug to the Cacti bug tracker
earlier last week and haven't had a response on it yet, though I see
responses have been posted for bugs newer than mine.

Just to confirm, are these bugs visible to developers? If so, shall I
leave it there, or would it be better to contact a developer directly?

-- 
PGP key:
https://www.sensepost.com/pgp/frank.txt

Solved the issue, 

For future reference, my script server data query inputs (|query_*|), for some hosts, didn't have unique indexes.
update_data_query_sort_cache -> get_ordered_index_type_list -> " /* fields that contain duplicate or empty values are not suitable to index off of */"
Since there were duplicates in all "|query_|" variables none was returned. 
I wasn't aware that those (just variables!) must be unique, for me the combination of two input fields was unique so I've added the combination of those two input fields as another query input field (otherwise unused at all in data sources or graph title names).

 


-----Original Message-----
From: Nitzan Zaifman 
Sent: Thursday, July 09, 2015 11:44 AM
To: Support requests/questions about cacti
Subject: Re: [cacti-user] Cacti data sources and graphs not created (complex)

Well great to see this list so active. The forums aren't, is thereanywhere to get someone's ear?...

I'm almost 100% positive that it's a real Cacti issue either in Cacti itself or the database (definition). How can I check this? Logs? PHP file? Etc.
There are many hosts that this dataquery script doesn't work on - the commonality is those hosts return 150 items 75 rows results.
There are hosts were it works and the same query returns 8 items, 4 rows there. (results count difference is correct, it depends on the job of those hosts).
I believe this is the only difference.

Eagerly awaiting reply...


-----Original Message-----
From: Nitzan Zaifman
Sent: Tuesday, July 07, 2015 9:39 AM
To: Support requests/questions about cacti
Subject: Re: [cacti-user] Cacti data sources and graphs not created (complex)

Well, the screenshot attachments didn't work...
I hope my description is good enough...

-----Original Message-----
From: Nitzan Zaifman
Sent: Tuesday, July 07, 2015 9:20 AM
To: Support requests/questions about cacti
Subject: [cacti-user] Cacti data sources and graphs not created (complex)

Hi,

General / summary:
One of my data queries doesn't work. It's a new one but the same one works on one host and not on another (actually doesn't work on a couple of others tried).
>From initial (as far as I could) analysis - the data query does not create the graph and data sources correctly on the effected host(/hosts) and rrd is not created (nor does poller ever retrieves the data).

********************************************************************************
How do I see a log of a new graph creation (from device -> data query), not the poller events (that aren't, there are none).
********************************************************************************

Specific:
I've took example of two hosts: "Peach" and "Orange", both trying to add the same generated data queries (from script server) graphs.
It all works well as expected in "Orange", but does not in "Peach"
I've included everything in screenshots (those are temporary internal lab servers so I don't mind) so you'd get the full picture.
The two differences I can see tell between the two are:

1.       "Peach"'s (not working) MVSnew dataquery is significantly larger (150 items, 75 rows) vs. "Orange"'s one (8 items, 4 rows).

2.       Peach has a lot of script server data queires, about 65, vs. very little on Orange.

Attachments:
General:
DataQueriesMvsnew.PNG - shows MVSnew dataquery configuration with its associated graph templates "Active" and "Rate".
DataQueriesMvsnewActive.PNG - shows for example "Active" datasource to graph mapping.

Peach (not working example):
CreatingANewGraphPeach.PNG - shows that indeed the data query works as expected and I get my variables ("MVS name" and "service").
PROBLEM-CreatingANewGraphPeachProblem.PNG - the new graph was not created correctly and did not resolve variables.
PROBLEM-NotCreatedDatasourcePeachOrange.PNG - the datasource was not created correctly, btw the rrd file did not create at all PROBLEM-NotCreatedPollerCacheEnteriesPeach.PNG - poller cache was not created for this new graph (Cacti's log only shows poller events, where do I see creation problems?)

Orange (working example):
CreatingANewGraphOrange.PNG - with the same exact MVSnew data query, I've created a graph on another host, Orange (here "Rate" example) CreatingANewGraphOrangeWORKS.PNG - graph created correctly and resolved variable names.
CreatedDatasourceOrange.PNG - datasource created correctly (including rrd file which was never generated in "Peach") CreatedPollerCacheEnteriesOrange.PNG - relevant poller cache created correctly.



------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
cacti-user mailing list
cac...@li...
https://lists.sourceforge.net/lists/listinfo/cacti-user

------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
cacti-user mailing list
cac...@li...
https://lists.sourceforge.net/lists/listinfo/cacti-user

Well great to see this list so active. The forums aren't, is thereanywhere to get someone's ear?...

I'm almost 100% positive that it's a real Cacti issue either in Cacti itself or the database (definition). How can I check this? Logs? PHP file? Etc.
There are many hosts that this dataquery script doesn't work on - the commonality is those hosts return 150 items 75 rows results.
There are hosts were it works and the same query returns 8 items, 4 rows there. (results count difference is correct, it depends on the job of those hosts).
I believe this is the only difference.

Eagerly awaiting reply...


-----Original Message-----
From: Nitzan Zaifman 
Sent: Tuesday, July 07, 2015 9:39 AM
To: Support requests/questions about cacti
Subject: Re: [cacti-user] Cacti data sources and graphs not created (complex)

Well, the screenshot attachments didn't work...
I hope my description is good enough...

-----Original Message-----
From: Nitzan Zaifman
Sent: Tuesday, July 07, 2015 9:20 AM
To: Support requests/questions about cacti
Subject: [cacti-user] Cacti data sources and graphs not created (complex)

Hi,

General / summary:
One of my data queries doesn't work. It's a new one but the same one works on one host and not on another (actually doesn't work on a couple of others tried).
>From initial (as far as I could) analysis - the data query does not create the graph and data sources correctly on the effected host(/hosts) and rrd is not created (nor does poller ever retrieves the data).

********************************************************************************
How do I see a log of a new graph creation (from device -> data query), not the poller events (that aren't, there are none).
********************************************************************************

Specific:
I've took example of two hosts: "Peach" and "Orange", both trying to add the same generated data queries (from script server) graphs.
It all works well as expected in "Orange", but does not in "Peach"
I've included everything in screenshots (those are temporary internal lab servers so I don't mind) so you'd get the full picture.
The two differences I can see tell between the two are:

1.       "Peach"'s (not working) MVSnew dataquery is significantly larger (150 items, 75 rows) vs. "Orange"'s one (8 items, 4 rows).

2.       Peach has a lot of script server data queires, about 65, vs. very little on Orange.

Attachments:
General:
DataQueriesMvsnew.PNG - shows MVSnew dataquery configuration with its associated graph templates "Active" and "Rate".
DataQueriesMvsnewActive.PNG - shows for example "Active" datasource to graph mapping.

Peach (not working example):
CreatingANewGraphPeach.PNG - shows that indeed the data query works as expected and I get my variables ("MVS name" and "service").
PROBLEM-CreatingANewGraphPeachProblem.PNG - the new graph was not created correctly and did not resolve variables.
PROBLEM-NotCreatedDatasourcePeachOrange.PNG - the datasource was not created correctly, btw the rrd file did not create at all PROBLEM-NotCreatedPollerCacheEnteriesPeach.PNG - poller cache was not created for this new graph (Cacti's log only shows poller events, where do I see creation problems?)

Orange (working example):
CreatingANewGraphOrange.PNG - with the same exact MVSnew data query, I've created a graph on another host, Orange (here "Rate" example) CreatingANewGraphOrangeWORKS.PNG - graph created correctly and resolved variable names.
CreatedDatasourceOrange.PNG - datasource created correctly (including rrd file which was never generated in "Peach") CreatedPollerCacheEnteriesOrange.PNG - relevant poller cache created correctly.



------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
cacti-user mailing list
cac...@li...
https://lists.sourceforge.net/lists/listinfo/cacti-user

Well, the screenshot attachments didn't work...
I hope my description is good enough...

-----Original Message-----
From: Nitzan Zaifman 
Sent: Tuesday, July 07, 2015 9:20 AM
To: Support requests/questions about cacti
Subject: [cacti-user] Cacti data sources and graphs not created (complex)

Hi,

General / summary:
One of my data queries doesn't work. It's a new one but the same one works on one host and not on another (actually doesn't work on a couple of others tried).
>From initial (as far as I could) analysis - the data query does not create the graph and data sources correctly on the effected host(/hosts) and rrd is not created (nor does poller ever retrieves the data).

********************************************************************************
How do I see a log of a new graph creation (from device -> data query), not the poller events (that aren't, there are none).
********************************************************************************

Specific:
I've took example of two hosts: "Peach" and "Orange", both trying to add the same generated data queries (from script server) graphs.
It all works well as expected in "Orange", but does not in "Peach"
I've included everything in screenshots (those are temporary internal lab servers so I don't mind) so you'd get the full picture.
The two differences I can see tell between the two are:

1.       "Peach"'s (not working) MVSnew dataquery is significantly larger (150 items, 75 rows) vs. "Orange"'s one (8 items, 4 rows).

2.       Peach has a lot of script server data queires, about 65, vs. very little on Orange.

Attachments:
General:
DataQueriesMvsnew.PNG - shows MVSnew dataquery configuration with its associated graph templates "Active" and "Rate".
DataQueriesMvsnewActive.PNG - shows for example "Active" datasource to graph mapping.

Peach (not working example):
CreatingANewGraphPeach.PNG - shows that indeed the data query works as expected and I get my variables ("MVS name" and "service").
PROBLEM-CreatingANewGraphPeachProblem.PNG - the new graph was not created correctly and did not resolve variables.
PROBLEM-NotCreatedDatasourcePeachOrange.PNG - the datasource was not created correctly, btw the rrd file did not create at all PROBLEM-NotCreatedPollerCacheEnteriesPeach.PNG - poller cache was not created for this new graph (Cacti's log only shows poller events, where do I see creation problems?)

Orange (working example):
CreatingANewGraphOrange.PNG - with the same exact MVSnew data query, I've created a graph on another host, Orange (here "Rate" example) CreatingANewGraphOrangeWORKS.PNG - graph created correctly and resolved variable names.
CreatedDatasourceOrange.PNG - datasource created correctly (including rrd file which was never generated in "Peach") CreatedPollerCacheEnteriesOrange.PNG - relevant poller cache created correctly.

Hi,

General / summary:
One of my data queries doesn't work. It's a new one but the same one works on one host and not on another (actually doesn't work on a couple of others tried).
>From initial (as far as I could) analysis - the data query does not create the graph and data sources correctly on the effected host(/hosts) and rrd is not created (nor does poller ever retrieves the data).

********************************************************************************
How do I see a log of a new graph creation (from device -> data query), not the poller events (that aren't, there are none).
********************************************************************************

Specific:
I've took example of two hosts: "Peach" and "Orange", both trying to add the same generated data queries (from script server) graphs.
It all works well as expected in "Orange", but does not in "Peach"
I've included everything in screenshots (those are temporary internal lab servers so I don't mind) so you'd get the full picture.
The two differences I can see tell between the two are:

1.       "Peach"'s (not working) MVSnew dataquery is significantly larger (150 items, 75 rows) vs. "Orange"'s one (8 items, 4 rows).

2.       Peach has a lot of script server data queires, about 65, vs. very little on Orange.

Attachments:
General:
DataQueriesMvsnew.PNG - shows MVSnew dataquery configuration with its associated graph templates "Active" and "Rate".
DataQueriesMvsnewActive.PNG - shows for example "Active" datasource to graph mapping.

Peach (not working example):
CreatingANewGraphPeach.PNG - shows that indeed the data query works as expected and I get my variables ("MVS name" and "service").
PROBLEM-CreatingANewGraphPeachProblem.PNG - the new graph was not created correctly and did not resolve variables.
PROBLEM-NotCreatedDatasourcePeachOrange.PNG - the datasource was not created correctly, btw the rrd file did not create at all
PROBLEM-NotCreatedPollerCacheEnteriesPeach.PNG - poller cache was not created for this new graph (Cacti's log only shows poller events, where do I see creation problems?)

Orange (working example):
CreatingANewGraphOrange.PNG - with the same exact MVSnew data query, I've created a graph on another host, Orange (here "Rate" example)
CreatingANewGraphOrangeWORKS.PNG - graph created correctly and resolved variable names.
CreatedDatasourceOrange.PNG - datasource created correctly (including rrd file which was never generated in "Peach")
CreatedPollerCacheEnteriesOrange.PNG - relevant poller cache created correctly.

There is IMHO a feature in latest spine to print runtime per host. Set
"--verbosity=3" to get that output (among a lot of other output). This way you
learn, which host is suffering most. In the cases I have looking at, this
narrows down and you will very likely know, which script is performing worst.

Another way is to do a "ps -ef" and see, which script is running as "cactiuser"
at the end of the run. This is not exact, as this script may have been waiting
on completion of a previous script in the same process.
When poller.php is working hard, this is very likely the "rrdtool update" phase,
which is the last one to be completed. Using BOOST is the way to tackle that (or
using ssd devices).

Regarding the "patch": With a bit of C knowledge, you may contribute this one.
Unfortunately, you will need a new database field to store that data ...

Reinhard


Am 01.07.2015 um 18:58 schrieb Greiner Martin:
> Okay, so I have to wait till this is patched in spine?!
> At the end of the "run" I can see the poller.php working hard,
> and two spine processes are running.
> Any hints how to get out which script took as much time?
> 
> Martin
> 
> 
> -----Ursprüngliche Nachricht-----
> Von: Gandalf [mailto:ga...@ca...] 
> Gesendet: Mittwoch, 01. Juli 2015 18:50
> An: Support requests/questions about cacti
> Betreff: [MASSMAIL]Re: [cacti-user] Performance
> 
> The "weight" is determinded by counting datasources. Each datasource is counted as "1".
> In case of a "pure SNMP only" system, this might be a quite good approach. But it is clear, that runtime of e.g. scripts will cause quite different runtimes of spine processes. And this is obviously what you see.
> To do better, this would need a per-datasource runtime counter instead of "1".
> Patches against spine to implement this are appreciated.
> 
> Reinhard
> 
> Am 01.07.2015 um 18:18 schrieb Greiner Martin:
>> Hi
>>
>> We are allready using spine,
>> but the balancing from hosts to processes/threads seems not to be the best.
>>
>> The first 20 seconds, the system will be under heavy load, but later 
>> only two spine processes are running and it will already take about 230 seconds to complete one run.
>>
>> Do you have any suggestions about optimization?
>>
>> Technical Support
>> General Information
>> Date      Wed, 01 Jul 2015 17:59:34 +0200
>> Cacti Version     0.8.8c
>> Cacti OS               unix
>> SNMP Version NET-SNMP version: 5.4.3
>> RRDTool Version             RRDTool 1.4.x
>> Hosts    1252
>> Graphs                 22991
>> Data Sources     Script/Command: 1467
>> SNMP: 4369
>> SNMP Query: 24725
>> Script Query: 73
>> Script - Script Server (PHP): 3
>> Script Query - Script Server: 575
>> Total: 31212
>> Poller Information
>> Interval                300
>> Type      SPINE 0.8.8c Copyright 2002-2014 by The Cacti Group
>> Items    Action[0]: 42417
>> Action[1]: 1696
>> Action[2]: 1169
>> Total: 45282
>> Concurrent Processes   4
>> Max Threads     16
>> PHP Servers       10
>> Script Timeout 25
>> Max OID              20
>> Last Run Statistics           Time:231.4831 Method:spine Processes:4 Threads:16 Hosts:1252 HostsPerProcess:313 DataSources:45282 RRDsProcessed:29851
>> PHP Information
>> PHP Version      5.3.3-7+squeeze23
>> PHP OS                Linux
>> PHP uname        Linux cacti.noc 3.16.0-vmware20150330-1-amd64 #1 SMP Debian 3.16.7-1~vmware20150330 (2015-03-30) x86_64
>> PHP SNMP         Installed
>> max_execution_time   30
>> memory_limit   512M
>>
>>
>> regards
>> Martin Greiner
>>
>>
>> ----------------------------------------------------------------------
>> -------- Don't Limit Your Business. Reach for the Cloud.
>> GigeNET's Cloud Solutions provide you with the tools and support that 
>> you need to offload your IT needs and focus on growing your business.
>> Configured For All Businesses. Start Your Cloud Today.
>> https://www.gigenetcloud.com/
>> _______________________________________________
>> cacti-user mailing list
>> cac...@li...
>> https://lists.sourceforge.net/lists/listinfo/cacti-user
>>
> 
> ------------------------------------------------------------------------------
> Don't Limit Your Business. Reach for the Cloud.
> GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business.
> Configured For All Businesses. Start Your Cloud Today.
> https://www.gigenetcloud.com/
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
> 
> ------------------------------------------------------------------------------
> Don't Limit Your Business. Reach for the Cloud.
> GigeNET's Cloud Solutions provide you with the tools and support that
> you need to offload your IT needs and focus on growing your business.
> Configured For All Businesses. Start Your Cloud Today.
> https://www.gigenetcloud.com/
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
> 

Okay, so I have to wait till this is patched in spine?!
At the end of the "run" I can see the poller.php working hard,
and two spine processes are running.
Any hints how to get out which script took as much time?

Martin


-----Ursprüngliche Nachricht-----
Von: Gandalf [mailto:ga...@ca...] 
Gesendet: Mittwoch, 01. Juli 2015 18:50
An: Support requests/questions about cacti
Betreff: [MASSMAIL]Re: [cacti-user] Performance

The "weight" is determinded by counting datasources. Each datasource is counted as "1".
In case of a "pure SNMP only" system, this might be a quite good approach. But it is clear, that runtime of e.g. scripts will cause quite different runtimes of spine processes. And this is obviously what you see.
To do better, this would need a per-datasource runtime counter instead of "1".
Patches against spine to implement this are appreciated.

Reinhard

Am 01.07.2015 um 18:18 schrieb Greiner Martin:
> Hi
> 
> We are allready using spine,
> but the balancing from hosts to processes/threads seems not to be the best.
> 
> The first 20 seconds, the system will be under heavy load, but later 
> only two spine processes are running and it will already take about 230 seconds to complete one run.
> 
> Do you have any suggestions about optimization?
> 
> Technical Support
> General Information
> Date      Wed, 01 Jul 2015 17:59:34 +0200
> Cacti Version     0.8.8c
> Cacti OS               unix
> SNMP Version NET-SNMP version: 5.4.3
> RRDTool Version             RRDTool 1.4.x
> Hosts    1252
> Graphs                 22991
> Data Sources     Script/Command: 1467
> SNMP: 4369
> SNMP Query: 24725
> Script Query: 73
> Script - Script Server (PHP): 3
> Script Query - Script Server: 575
> Total: 31212
> Poller Information
> Interval                300
> Type      SPINE 0.8.8c Copyright 2002-2014 by The Cacti Group
> Items    Action[0]: 42417
> Action[1]: 1696
> Action[2]: 1169
> Total: 45282
> Concurrent Processes   4
> Max Threads     16
> PHP Servers       10
> Script Timeout 25
> Max OID              20
> Last Run Statistics           Time:231.4831 Method:spine Processes:4 Threads:16 Hosts:1252 HostsPerProcess:313 DataSources:45282 RRDsProcessed:29851
> PHP Information
> PHP Version      5.3.3-7+squeeze23
> PHP OS                Linux
> PHP uname        Linux cacti.noc 3.16.0-vmware20150330-1-amd64 #1 SMP Debian 3.16.7-1~vmware20150330 (2015-03-30) x86_64
> PHP SNMP         Installed
> max_execution_time   30
> memory_limit   512M
> 
> 
> regards
> Martin Greiner
> 
> 
> ----------------------------------------------------------------------
> -------- Don't Limit Your Business. Reach for the Cloud.
> GigeNET's Cloud Solutions provide you with the tools and support that 
> you need to offload your IT needs and focus on growing your business.
> Configured For All Businesses. Start Your Cloud Today.
> https://www.gigenetcloud.com/
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
> 

------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
cacti-user mailing list
cac...@li...
https://lists.sourceforge.net/lists/listinfo/cacti-user

The "weight" is determinded by counting datasources. Each datasource is counted
as "1".
In case of a "pure SNMP only" system, this might be a quite good approach. But
it is clear, that runtime of e.g. scripts will cause quite different runtimes of
spine processes. And this is obviously what you see.
To do better, this would need a per-datasource runtime counter instead of "1".
Patches against spine to implement this are appreciated.

Reinhard

Am 01.07.2015 um 18:18 schrieb Greiner Martin:
> Hi
> 
> We are allready using spine,
> but the balancing from hosts to processes/threads seems not to be the best.
> 
> The first 20 seconds, the system will be under heavy load,
> but later only two spine processes are running and it will already take about 230 seconds to complete one run.
> 
> Do you have any suggestions about optimization?
> 
> Technical Support
> General Information
> Date      Wed, 01 Jul 2015 17:59:34 +0200
> Cacti Version     0.8.8c
> Cacti OS               unix
> SNMP Version NET-SNMP version: 5.4.3
> RRDTool Version             RRDTool 1.4.x
> Hosts    1252
> Graphs                 22991
> Data Sources     Script/Command: 1467
> SNMP: 4369
> SNMP Query: 24725
> Script Query: 73
> Script - Script Server (PHP): 3
> Script Query - Script Server: 575
> Total: 31212
> Poller Information
> Interval                300
> Type      SPINE 0.8.8c Copyright 2002-2014 by The Cacti Group
> Items    Action[0]: 42417
> Action[1]: 1696
> Action[2]: 1169
> Total: 45282
> Concurrent Processes   4
> Max Threads     16
> PHP Servers       10
> Script Timeout 25
> Max OID              20
> Last Run Statistics           Time:231.4831 Method:spine Processes:4 Threads:16 Hosts:1252 HostsPerProcess:313 DataSources:45282 RRDsProcessed:29851
> PHP Information
> PHP Version      5.3.3-7+squeeze23
> PHP OS                Linux
> PHP uname        Linux cacti.noc 3.16.0-vmware20150330-1-amd64 #1 SMP Debian 3.16.7-1~vmware20150330 (2015-03-30) x86_64
> PHP SNMP         Installed
> max_execution_time   30
> memory_limit   512M
> 
> 
> regards
> Martin Greiner
> 
> 
> ------------------------------------------------------------------------------
> Don't Limit Your Business. Reach for the Cloud.
> GigeNET's Cloud Solutions provide you with the tools and support that
> you need to offload your IT needs and focus on growing your business.
> Configured For All Businesses. Start Your Cloud Today.
> https://www.gigenetcloud.com/
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
> 

Hi

We are allready using spine,
but the balancing from hosts to processes/threads seems not to be the best.

The first 20 seconds, the system will be under heavy load,
but later only two spine processes are running and it will already take about 230 seconds to complete one run.

Do you have any suggestions about optimization?

Technical Support
General Information
Date      Wed, 01 Jul 2015 17:59:34 +0200
Cacti Version     0.8.8c
Cacti OS               unix
SNMP Version NET-SNMP version: 5.4.3
RRDTool Version             RRDTool 1.4.x
Hosts    1252
Graphs                 22991
Data Sources     Script/Command: 1467
SNMP: 4369
SNMP Query: 24725
Script Query: 73
Script - Script Server (PHP): 3
Script Query - Script Server: 575
Total: 31212
Poller Information
Interval                300
Type      SPINE 0.8.8c Copyright 2002-2014 by The Cacti Group
Items    Action[0]: 42417
Action[1]: 1696
Action[2]: 1169
Total: 45282
Concurrent Processes   4
Max Threads     16
PHP Servers       10
Script Timeout 25
Max OID              20
Last Run Statistics           Time:231.4831 Method:spine Processes:4 Threads:16 Hosts:1252 HostsPerProcess:313 DataSources:45282 RRDsProcessed:29851
PHP Information
PHP Version      5.3.3-7+squeeze23
PHP OS                Linux
PHP uname        Linux cacti.noc 3.16.0-vmware20150330-1-amd64 #1 SMP Debian 3.16.7-1~vmware20150330 (2015-03-30) x86_64
PHP SNMP         Installed
max_execution_time   30
memory_limit   512M


regards
Martin Greiner

Thank you very much again... 
I placed it here so you understand exactly why I asked this, but you got it straight away anyhow.

Best regards.



-----Original Message-----
From: Gandalf [mailto:ga...@ca...] 
Sent: Thursday, June 18, 2015 9:57 PM
To: Support requests/questions about cacti
Subject: Re: [cacti-user] Cacti script_server index invocation



Am 18.06.2015 um 20:37 schrieb Nitzan Zaifman:
> Thanks you very much,
> 
> Quote:
> '
> "Index Count Changed" == Cacti _always_ gets the number of indexes '
> 
> Does this mean that cacti_always_gets the index count and if I happen to choose: "index count changed" then a full re-index will occur if that number is changed?

No, Cacti gets the index count only, when "Index Count Changed" has been chosen (or on a full reindex; e.g. when you manually perform a reindex operation).

> Or it means that ONLY had I choose this "index count changed" method, then *and only then* will cacti_always_gets this num_indexes? If I choose something else, then cacti won't get num_indexes nor the whole indexes.
Yes, that's it

> In <some of my> scripts the cost of "index" and "num_index" is the same and very high compared to specific data asked. I want to avoid num_indexes and index queries as much as possible.

Yes, that may happen. In this case, only you may decide if a faster way of "counting" exists. If it does not exist, it is _not_ recommended to select that reiondex option without need. But if "Uptime goes backwards" is the wrong solution, you may instead want to schedule a e.g. daily reindex on those data queries.

> Can data queries (<input>) generate index queries?
> 
> I use script server scripts as a mean to "cache" results of other scripts, with the add benefit of allowing GUI access to choose interested stats with "queries", the data itself (results of other scripts) is saved in static array variables thus filled only once and read for whatever's needed. 
> Some of those underlying scripts can be very expensive. 
> BTW, IMHO this technique is very elegant and affective for those kind of things and should go to some FAQ or a new manual in the future, but NM that now.
Please add such a topic to the forums or to the documentation, if need be.
I already did a very similar approach in the past but it seemed to me to be very specific. Perhaps you have found a better approach.


> BTW2, I use the same technique on full PHP script_server scripts (i.e. not running underlying scripts), and there I can granularly add stuff to my static data structures on first access - if it interests anyone on the list I'll provide an example.
> 
> Sorry for the lengthy mail  and hope that that clarifies what I'm after.
> If you need an example, this is part of a script server that runs expansive underlying scripts (the relevant cmd part):
> switch ($cmd) {
This better fits into the forums. There's not that much traffic on the mailing list Reinhard

------------------------------------------------------------------------------
_______________________________________________
cacti-user mailing list
cac...@li...
https://lists.sourceforge.net/lists/listinfo/cacti-user

Am 18.06.2015 um 20:37 schrieb Nitzan Zaifman:
> Thanks you very much, 
> 
> Quote:
> '
> "Index Count Changed" == Cacti _always_ gets the number of indexes
> '
> 
> Does this mean that cacti_always_gets the index count and if I happen to choose: "index count changed" then a full re-index will occur if that number is changed?

No, Cacti gets the index count only, when "Index Count Changed" has been chosen
(or on a full reindex; e.g. when you manually perform a reindex operation).

> Or it means that ONLY had I choose this "index count changed" method, then *and only then* will cacti_always_gets this num_indexes? If I choose something else, then cacti won't get num_indexes nor the whole indexes.
Yes, that's it

> In <some of my> scripts the cost of "index" and "num_index" is the same and very high compared to specific data asked. I want to avoid num_indexes and index queries as much as possible.

Yes, that may happen. In this case, only you may decide if a faster way of
"counting" exists. If it does not exist, it is _not_ recommended to select that
reiondex option without need. But if "Uptime goes backwards" is the wrong
solution, you may instead want to schedule a e.g. daily reindex on those data
queries.

> Can data queries (<input>) generate index queries?
> 
> I use script server scripts as a mean to "cache" results of other scripts, with the add benefit of allowing GUI access to choose interested stats with "queries", the data itself (results of other scripts) is saved in static array variables thus filled only once and read for whatever's needed. 
> Some of those underlying scripts can be very expensive. 
> BTW, IMHO this technique is very elegant and affective for those kind of things and should go to some FAQ or a new manual in the future, but NM that now.
Please add such a topic to the forums or to the documentation, if need be.
I already did a very similar approach in the past but it seemed to me to be very
specific. Perhaps you have found a better approach.


> BTW2, I use the same technique on full PHP script_server scripts (i.e. not running underlying scripts), and there I can granularly add stuff to my static data structures on first access - if it interests anyone on the list I'll provide an example.
> 
> Sorry for the lengthy mail  and hope that that clarifies what I'm after.
> If you need an example, this is part of a script server that runs expansive underlying scripts (the relevant cmd part):
> switch ($cmd) {
This better fits into the forums. There's not that much traffic on the mailing list
Reinhard

Thanks you very much, 

Quote:
'
"Index Count Changed" == Cacti _always_ gets the number of indexes
'

Does this mean that cacti_always_gets the index count and if I happen to choose: "index count changed" then a full re-index will occur if that number is changed?
Or it means that ONLY had I choose this "index count changed" method, then *and only then* will cacti_always_gets this num_indexes? If I choose something else, then cacti won't get num_indexes nor the whole indexes.
In <some of my> scripts the cost of "index" and "num_index" is the same and very high compared to specific data asked. I want to avoid num_indexes and index queries as much as possible.
Can data queries (<input>) generate index queries?

I use script server scripts as a mean to "cache" results of other scripts, with the add benefit of allowing GUI access to choose interested stats with "queries", the data itself (results of other scripts) is saved in static array variables thus filled only once and read for whatever's needed. 
Some of those underlying scripts can be very expensive. 
BTW, IMHO this technique is very elegant and affective for those kind of things and should go to some FAQ or a new manual in the future, but NM that now.
BTW2, I use the same technique on full PHP script_server scripts (i.e. not running underlying scripts), and there I can granularly add stuff to my static data structures on first access - if it interests anyone on the list I'll provide an example.

Sorry for the lengthy mail  and hope that that clarifies what I'm after.
If you need an example, this is part of a script server that runs expansive underlying scripts (the relevant cmd part):
switch ($cmd) {
        case "index":
       /*************/
            # we have to have the whole bunch for index, I hope Cacti is smart enough not to ask this too much (only when asked with run query)... :/
            if (!isset($statsArr[$host][$pythonScript])) {
                $statsArr[$host][$pythonScript] = ss_runPythonCmd_init($pythonScript, $host);
            }
            foreach (array_keys($statsArr[$host][$pythonScript]) as $key) {
                $returnStr .= $key.PHP_EOL;
            }
            break;

        case "num_indexes":
       /*******************/
            # we have to have the whole bunch for num_indexes, I hope Cacti is smart enough not to ask this too much (only when asked with run query)... :/
            if (!isset($statsArr[$host][$pythonScript])) {
                $statsArr[$host][$pythonScript] = ss_runPythonCmd_init($pythonScript, $host);
            }
            $returnStr = count($statsArr[$host][$pythonScript]).PHP_EOL;
            break;

        case "query":
       /*************/
            # we have to have the whole bunch for query, I hope Cacti is smart enough not to ask this too much (i.e. not every poll, only when asked with run query)... :/
            if (!isset($statsArr[$host][$pythonScript])) {
                $statsArr[$host][$pythonScript] = ss_runPythonCmd_init($pythonScript, $host);
            }

            # support asking for services and service1, service2, ... queries (Cacti var |query_service2| for ex.)
            foreach (array_keys($statsArr[$host][$pythonScript]) as $key) {
                if ($property == 'services') {
                    $returnStr .= $key.':'.$key.PHP_EOL;
                } elseif (preg_match("/^service(\d+)$/", $property, $matches)) {
                    $serviceArr=explode('.', $key);
                    $returnStr .= $key.':'.$serviceArr[$matches[1]-1].PHP_EOL;
                } else {
                    $returnStr .= $key.':'.$statsArr[$host][$pythonScript][$key][$property].PHP_EOL;
                }
            }
            break;

        case "get":
       /***********/

....



Thank you very much again.



-----Original Message-----
From: Gandalf [mailto:ga...@ca...] 
Sent: Thursday, June 18, 2015 8:06 PM
To: Support requests/questions about cacti
Subject: Re: [cacti-user] Cacti script_server index invocation

This depends on the "re-index setting".
"Uptime goes backwards" == Cacti tracks system uptime. If current value is lower than resent value (remembered by Cacti), Cacti will assume a rebbot and perform a reindex operation.
"Index Count Changed" == Cacti _always_ gets the number of indexes present (this is _not_ a full reindex operation). If previous count <> current count, a full reindex is done

Counting the index is done the following way:
- in case there is an OID that represents the index count, this single OID is fetched
- in case of a (server) script query, the script is called with the "index"
operator. Depending on the script, the related action is performed

Reinhard

Am 10.06.2015 um 11:30 schrieb Nitzan Zaifman:
> Hi,
> 
> I've tried the same question in the forums (http://forums.cacti.net/viewtopic.php?f=21&t=54713) for the last couple of days with no reply, I hope someone can pick it up here.
> 
> I don't recall encountering this in the docs nor the FAQs, if it's there please enlighten me and if not IMHO it belongs there.
> 
> I have a very expensive script server "index" query, one I want to do as seldom as possible.
> When does an index query called? once? whenever num_indexes change? can I set it up to only manually as num_indexes is equally as expensive.
> 
> Thanks, Bye, Nitzan.
> 
> ----------------------------------------------------------------------
> -------- _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
> 

------------------------------------------------------------------------------
_______________________________________________
cacti-user mailing list
cac...@li...
https://lists.sourceforge.net/lists/listinfo/cacti-user

 
 
************************************************************************************
This footnote confirms that this email message has been scanned by PineApp Mail-SeCure for the presence of malicious code, vandals & computer viruses.
************************************************************************************

This depends on the "re-index setting".
"Uptime goes backwards" == Cacti tracks system uptime. If current value is lower
than resent value (remembered by Cacti), Cacti will assume a rebbot and perform
a reindex operation.
"Index Count Changed" == Cacti _always_ gets the number of indexes present (this
is _not_ a full reindex operation). If previous count <> current count, a full
reindex is done

Counting the index is done the following way:
- in case there is an OID that represents the index count, this single OID is
fetched
- in case of a (server) script query, the script is called with the "index"
operator. Depending on the script, the related action is performed

Reinhard

Am 10.06.2015 um 11:30 schrieb Nitzan Zaifman:
> Hi,
> 
> I've tried the same question in the forums (http://forums.cacti.net/viewtopic.php?f=21&t=54713) for the last couple of days with no reply, I hope someone can pick it up here.
> 
> I don't recall encountering this in the docs nor the FAQs, if it's there please enlighten me and if not IMHO it belongs there.
> 
> I have a very expensive script server "index" query, one I want to do as seldom as possible.
> When does an index query called? once? whenever num_indexes change? can I set it up to only manually as num_indexes is equally as expensive.
> 
> Thanks, Bye, Nitzan.
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user
> 

Better yet/asked: 

Also when does a "query" is being done?

-----Original Message-----
From: Nitzan Zaifman 
Sent: Wednesday, June 10, 2015 12:31 PM
To: cac...@li...
Subject: [cacti-user] Cacti script_server index invocation

Hi,

I've tried the same question in the forums (http://forums.cacti.net/viewtopic.php?f=21&t=54713) for the last couple of days with no reply, I hope someone can pick it up here.

I don't recall encountering this in the docs nor the FAQs, if it's there please enlighten me and if not IMHO it belongs there.

I have a very expensive script server "index" query, one I want to do as seldom as possible.
When does an index query called? once? whenever num_indexes change? can I set it up to only manually as num_indexes is equally as expensive.

Thanks, Bye, Nitzan.

------------------------------------------------------------------------------
_______________________________________________
cacti-user mailing list
cac...@li...
https://lists.sourceforge.net/lists/listinfo/cacti-user

Hi,

I've tried the same question in the forums (http://forums.cacti.net/viewtopic.php?f=21&t=54713) for the last couple of days with no reply, I hope someone can pick it up here.

I don't recall encountering this in the docs nor the FAQs, if it's there please enlighten me and if not IMHO it belongs there.

I have a very expensive script server "index" query, one I want to do as seldom as possible.
When does an index query called? once? whenever num_indexes change? can I set it up to only manually as num_indexes is equally as expensive.

Thanks, Bye, Nitzan.

Le 2015-05-27 05:08, 宮崎昌男 a écrit :
> What is the meaning of "Limit"?
> Does the "Limit" meanY the maximum value for a network?
> Say, a 10G network, if I set the template for "Upper Limit",
> so I can get 10G graph.
> Is that correct?

Yes, the units displayed vertically will end with the value specified in 
upper limit. So if you use 10000000 as upper limit, you will get 10G as 
max value for the graph.

 From the rrdtool manual:

"Limits

[-u|--upper-limit value] [-l|--lower-limit value] [-r|--rigid]

By default the graph will be autoscaling so that it will adjust the 
y-axis to the range of the data. You can change this behavior by 
explicitly setting the limits. The displayed y-axis will then range at 
least from lower-limit to upper-limit. Autoscaling will still permit 
those boundaries to be stretched unless the rigid option is set."

http://oss.oetiker.ch/rrdtool/doc/rrdgraph.en.html#ILimits

---
Vincent Lequertier
sk...@ri...


> ------------------------------------------------------------------------------
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user

Hi. Vincent.
Thanks for the great help and for the quick answer.

The answer maybe just what I need.
But I have another question to ask.
What is the meaning of "Limit"?
Does the "Limit" mean the maximum value for a network?
Say, a 10G network, if I set the template for "Upper Limit",
so I can get 10G graph.
Is that correct?

Appreciate for any information.

cacti-user
05/27/2015

I'm not sure if I understand your question properly. If you wanna set a 
maximum value on the graph, you should tweak the graph template 
associated with the graph.

http://www.cacti.net/downloads/docs/html/graph_templates.html.

Unticking the "Auto Scale" option and setting the "Upper Limit" should 
give you what you want :-)

---
Vincent Lequertier
sk...@ri...

Le 2015-05-26 10:08, 宮崎昌男 a écrit :
> Hi.
> 
> We have a 10 Gigabit network. For the maximum value,
> we need 2G (2000M) to display on the graph data.
> Is there anyone know, it is possible? Or need more skill to customize?
> 
> OS: Redhat 7
> Cacti Verion:  cacti-0.8.8c
> 
> Appreciate for any help or advice!
> 
> 
> 
> Email: mmiyazaki@g.inetrely.co.jp
> 
> 
> cacti-user
> 05/25/2015
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across 
> Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable 
> Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> cacti-user mailing list
> cac...@li...
> https://lists.sourceforge.net/lists/listinfo/cacti-user

Hi.

We have a 10 Gigabit network. For the maximum value,
we need 2G (2000M) to display on the graph data.
Is there anyone know, it is possible? Or need more skill to customize?

OS: Redhat 7
Cacti Verion:  cacti-0.8.8c

Appreciate for any help or advice!



Email: mmiyazaki@g.inetrely.co.jp


cacti-user
05/25/2015

The issue I faced was with socket file. 
Fix1: configure mysql  
to have socket file as /var/run/mysql/mysqld.sock.
Fix2: Add following to php.ini:
mysql.default_socket=/path/to/mysql.sock 
- restart apache.
Here is my troubleshooting step: http://goo.gl/n5fEqh

Cheers.

Hi there Cacti Users,

I want to script the ability to add links to the weathermap.  Is there a 
tool out there that does this already?  Any ideas about how to make this 
task easier to do?

Cheers,

Noah

hello

I understand that I can enable vew to plugins or graph etc from
utilities -> user management -> select username -> realm permissions
but it has to be done one by one
I have hundred of users, I want all of them to view the plugin 
wheathermap tab
How can I enable / allow that view without seting it users by users .

Thanks .