Graphing Physical/Paging memory from Insight Manager
Moderators: Developers, Moderators
Graphing Physical/Paging memory from Insight Manager
Hello
I have written an ugly but potentially useful collection of scripts for graphing the cpqHoMemory branch from HP Insight Manager Agents.
All data can be found under .1.3.6.1.4.1.232.11.2.13 on any Proliant (or similar) system running windows with the Insight foundation agents.
Since people have been complaining about how difficult it is to find memory stats in windows machines I guess I can put them up here.
I have never exported templates before so the file may well turn your Cacti installation into a gibbering wreck, use with caution... =)
Cheers!
I have written an ugly but potentially useful collection of scripts for graphing the cpqHoMemory branch from HP Insight Manager Agents.
All data can be found under .1.3.6.1.4.1.232.11.2.13 on any Proliant (or similar) system running windows with the Insight foundation agents.
Since people have been complaining about how difficult it is to find memory stats in windows machines I guess I can put them up here.
I have never exported templates before so the file may well turn your Cacti installation into a gibbering wreck, use with caution... =)
Cheers!
- Attachments
-
- HP_insight_mem.xml
- His is my exported graph template with dependent stuff.
- (26.66 KiB) Downloaded 1904 times
-
- example of my memory graphs
- memgraph.JPG (46.93 KiB) Viewed 24653 times
My question
I added this templates to my cacti system,but found the graph is not like the author's graph.
Could some tell me why !
Thanks!
Could some tell me why !
Thanks!
- Attachments
-
- cacti_error_1.png (20.7 KiB) Viewed 24522 times
Fedora 3 | httpd-2.0.52-3 | php-4.3.9-3 | mysql-3.23.58-13 | net-snmp-5.1.2-11 | rrdtool-1.2.11 | cacti-0.8.6g
Dont know about the PHP errors, could be a version thing.
The graph and sources are fairly trivial to set up manually using Get SNMP Data as the input method for each of the four values:
cpqHoPhysicalMemorySize / cpqHoPhysicalMemoryFree
cpqHoPagingMemorySize / cpqHoPagingMemoryFree
First I convert the values from free to used, since that is more useful.
Then, in order to get the graph centered on the 0 line I negate all values for physical memory.
Could be that the CDEFs did not survive the traumatic export-import, so here they are for each source (in order represented in the graph):
cpqHoPagingMemorySize => No change
cpqHoPagingMemoryFree => "invert memfree" (dcef=a,b,-)
cpqHoPhysicalMemorySize => "negate" (cdef=CURRENT_DATA_SOURCE,-1,*)
cpqHoPhysicalMemoryFree => "physmem used negated" (cdef=c,d,-,-1,*)
Dont Negate the GPRINT stuff unless you like seeing negative printed values for RAM...
I'll be the first to admit that the naming standard and syntax for the cdefs should earn me a day at the spanking pole, feel free to improve it =)
Cheers!
The graph and sources are fairly trivial to set up manually using Get SNMP Data as the input method for each of the four values:
cpqHoPhysicalMemorySize / cpqHoPhysicalMemoryFree
cpqHoPagingMemorySize / cpqHoPagingMemoryFree
First I convert the values from free to used, since that is more useful.
Then, in order to get the graph centered on the 0 line I negate all values for physical memory.
Could be that the CDEFs did not survive the traumatic export-import, so here they are for each source (in order represented in the graph):
cpqHoPagingMemorySize => No change
cpqHoPagingMemoryFree => "invert memfree" (dcef=a,b,-)
cpqHoPhysicalMemorySize => "negate" (cdef=CURRENT_DATA_SOURCE,-1,*)
cpqHoPhysicalMemoryFree => "physmem used negated" (cdef=c,d,-,-1,*)
Dont Negate the GPRINT stuff unless you like seeing negative printed values for RAM...
I'll be the first to admit that the naming standard and syntax for the cdefs should earn me a day at the spanking pole, feel free to improve it =)
Cheers!
Just an FYI for anyone using this script: when I imported it on my server, the "physmem used negated" CDEF was incorrect which was causing my PHP segmentation fault errors. To fix this, edit the CDEF so that it looks like this:
When it imported, the "physmem used" was not selected as the CDEF to use and it was sitting blank. Cacti was either using nothing which caused an error in RRDTool OR (in my case) it referenced itself which put Cacti into an endless loop. Because I had a PHP segmentation fault just viewing the CDEF, I had to manually go in the database and correct the CDEF.Item #1 Another CDEF: physmem used
Item #2 Custom String: -1
Item #3 Operator: *
On the graph itself the RAM usage will be shown as negative (see the green values in the original post) because it's being inversed. The GPRINT stuff at the bottom of the graph should still be positive values.
After I fixed the CDEF, my graphs look exactly like the the ones in the original post. Frater, thank you for such a great template!
After I fixed the CDEF, my graphs look exactly like the the ones in the original post. Frater, thank you for such a great template!
Who is online
Users browsing this forum: No registered users and 4 guests