hi all, i push a lot of traffic on this one box. using the "interface statistics" graphs, i get the attached graph. snmpwalk on the oid shows a COUNTER32. i'm guessing that i'm getting non-smooth graphs due to wrapping occurring faster than the 300 second polling cycle. if so, what options do i have at my disposal?
thanks,
jlg
IF-MIB::ifInOctets.2 = Counter32: 1629128164
IF-MIB::ifInOctets.3 = Counter32: 259009177
screwy graphy -- counter cycling too fast?
Moderators: Developers, Moderators
screwy graphy -- counter cycling too fast?
- Attachments
-
- screwy-graph.jpg (47.13 KiB) Viewed 2668 times
ok, i found where to set the maximum value. it can be set by going to the "Data Source" for the graph.
to help others who stumble on this thread...it appears that there is some magic going on. if you examine the data template "Interface - Traffic", notice that Maximum Value has the "Use Per-Data Source Value (Ignore this Value)" check box checked. i would have expected this to prompt me for the value when creating the graph. however, it did not and appears to have used the value from the template (or maybe it was supposed to get it from the snmp query?).
anyway, i have set the value and am hoping it fixes the problem. looks like making the change doesn't fix existing data so i'll have to watch this for a day or so.
jlg
to help others who stumble on this thread...it appears that there is some magic going on. if you examine the data template "Interface - Traffic", notice that Maximum Value has the "Use Per-Data Source Value (Ignore this Value)" check box checked. i would have expected this to prompt me for the value when creating the graph. however, it did not and appears to have used the value from the template (or maybe it was supposed to get it from the snmp query?).
anyway, i have set the value and am hoping it fixes the problem. looks like making the change doesn't fix existing data so i'll have to watch this for a day or so.
jlg
thank egironda.
rrdtool info now shows the correct maximum but the old graph data remains broken. i'm assuming that this change will apply for new data...not old.
a couple of questions:
1. any good tips to avoid this issue going forward? i'm never given the opportunity to set the maximum as i create the graph (data source) and apparently changing the gui interface values after creation do not change the rrd file. i'd prefer to not have to rrdtune every file if avoidable.
2. is there a way to get the system to use the value from the xml query? i get "Error: Save Failed: Field Input Error (Check Red Fields)" when trying to use "|ifspeed|" or "|query_ifspeed|" for Maximum Value for traffic_in/out for the Data Template "Interface - Traffic".
thanks,
jlg
rrdtool info now shows the correct maximum but the old graph data remains broken. i'm assuming that this change will apply for new data...not old.
a couple of questions:
1. any good tips to avoid this issue going forward? i'm never given the opportunity to set the maximum as i create the graph (data source) and apparently changing the gui interface values after creation do not change the rrd file. i'd prefer to not have to rrdtune every file if avoidable.
2. is there a way to get the system to use the value from the xml query? i get "Error: Save Failed: Field Input Error (Check Red Fields)" when trying to use "|ifspeed|" or "|query_ifspeed|" for Maximum Value for traffic_in/out for the Data Template "Interface - Traffic".
thanks,
jlg
- TheWitness
- Developer
- Posts: 17007
- Joined: Tue May 14, 2002 5:08 pm
- Location: MI, USA
- Contact:
After you figure out about the 64bit counters, you should just blow away the old graph. Many servers have gig interfaces, but don't support the 64bit counters. Most network switches do support them though.
If you end of having to monitor one of those nasty servers, you should consider a different polling interval with a revised interface template to accomodate a, say 1 minute polling interval.
We have a non-published template that allows for 1 minute polling without polling "everything" at 1 minute. You need to send an e-mail to get it though.
TheWitness
If you end of having to monitor one of those nasty servers, you should consider a different polling interval with a revised interface template to accomodate a, say 1 minute polling interval.
We have a non-published template that allows for 1 minute polling without polling "everything" at 1 minute. You need to send an e-mail to get it though.
TheWitness
True understanding begins only when we realize how little we truly understand...
Life is an adventure, let yours begin with Cacti!
Author of dozens of Cacti plugins and customization's. Advocate of LAMP, MariaDB, IBM Spectrum LSF and the world of batch. Creator of IBM Spectrum RTM, author of quite a bit of unpublished work and most of Cacti's bugs.
_________________
Official Cacti Documentation
GitHub Repository with Supported Plugins
Percona Device Packages (no support)
Interesting Device Packages
For those wondering, I'm still here, but lost in the shadows. Yearning for less bugs. Who want's a Cacti 1.3/2.0? Streams anyone?
Life is an adventure, let yours begin with Cacti!
Author of dozens of Cacti plugins and customization's. Advocate of LAMP, MariaDB, IBM Spectrum LSF and the world of batch. Creator of IBM Spectrum RTM, author of quite a bit of unpublished work and most of Cacti's bugs.
_________________
Official Cacti Documentation
GitHub Repository with Supported Plugins
Percona Device Packages (no support)
Interesting Device Packages
For those wondering, I'm still here, but lost in the shadows. Yearning for less bugs. Who want's a Cacti 1.3/2.0? Streams anyone?
hi all, here's some follow up...
1. after having reset the maximum in the cacti gui and having used rrdtool to tune the maximum, my graphs still show the same broken behavior.
2. the snmp daemon used here is from net-snmp. i see "Counter32" from snmpget on the OID in question so it looks like 64 bit counters are not going to be available to me in the near term. so i'll ask TheWitness for the 1 minute template in email.
thanks,
jlg
1. after having reset the maximum in the cacti gui and having used rrdtool to tune the maximum, my graphs still show the same broken behavior.
2. the snmp daemon used here is from net-snmp. i see "Counter32" from snmpget on the OID in question so it looks like 64 bit counters are not going to be available to me in the near term. so i'll ask TheWitness for the 1 minute template in email.
thanks,
jlg
Who is online
Users browsing this forum: No registered users and 1 guest