Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

RHEL7 & NET-SNMP Memory Monitoring

Is anyone monitoring RHEL7 boxes with NET-SNMP and if so are you getting correct or even normal looking data for Memory Used/Memory Available?(Not Volume statistics) I recently noticed that all of our RHEL7 boxes are showing 0% for Memory Used/Available and when I look further at the box or pull it up in Perfstack the values are very much negative. Going through the SolarWinds formula for calculating NET-SNMP Memory also comes back negative. Yet when I talk to support they keep telling me that RHEL7 should work just fine for SNMP polling, but then suggestion I try the agent. So is anyone using net-snmp and seeing accurate numbers? I just want to know if it's something on our side with a configuraiton or if this is a known issue. I can't seem to get a solid confirmed answer from support and they just keep sending me links to documents that I'm well aware of....

Find more posts tagged with

Accepted answers

All comments

vinay.by

Do you know which OID Orion is using for SNMP Polls ? Have you tried checking if it has something to do with the SNMP OID that is being used ?

vinay.by

Why don't you test if for yourself, I am sure Poller Checker would still exist on your SolarWinds Primary Poller

How to Use Poller Checker Tool

vinay.by

How to Use Poller Checker

Success Center

How to check and Update the OIDs

Success Center

nglynn

I only know what support has told me which is the below OIDs that are used and how Orion calculates it. Now I won't even get into what my Linux Admins think about this formula..

Came from the SNMPwalk:
Total RAM in machine: .1.3.6.1.4.1.2021.4.5.0 = INTEGER: 197553772
Total RAM available: .1.3.6.1.4.1.2021.4.6.0 = INTEGER: 193026460
Total RAM Buffered: .1.3.6.1.4.1.2021.4.14.0 = INTEGER: 48
Total Cached Memory: .1.3.6.1.4.1.2021.4.15.0 = INTEGER: 190284140

1. To get the total RAM used:
Total RAM in machine - Total RAM Available = Total RAM Used 197553772 - 193026460 = 4527312

2. To get the Used Memory:
Total RAM Used - Cached Memory - RAM Buffered = Used Memory 4527312 - 190284140 - 48 = -185756876

3. To get the percentage:
(Used Memory / Total RAM) * 100 = -185756876 / 4527312 = -41.03027933572946 %

4527312 * 1024 = 4635967488 Total Ram
-185756876 * 1024 = -190215041024 Used Memory

Worth noting I'm not seeing issues on RHEL5 or RHEL6 boxes being polled in theory the same way using net-snmp.

vinay.by

Give it a try, just verify if its the same OID that Orion is using, the links that i pasted will help you in verifying the same.

If you have an old OID that is relevant you can as well modify that and force Orion to use that OID, Hope it helps

mesverrum

When Linux admins complain about the way net-snmp present memory I just use this, Linux Memory Monitoring Perl, SAM script monitor then we can just disable snmp based cpu/mem polling. That fight has been going on for yearrrrrs and is present across every snmp based monitoring platform. A quick calculation from net-snmp never reflects what the admin is really looking for so why fight them over it. Just give them the metrics that they want in the format they want them. Even when the numbers aren't completely wrong they are still subtly wrong or not showing the fine details they need to assess the server.

stevenstadel

We are having the exact same issue. Opened support request today.

Gauges are displaying Memory at 0%, and the Average Memory Usage Graph is showing massive negative numbers.

We have checked the OIDs and they look like they are pulling the correct values.

Doing a spot check of our Linux nodes show some of these have been displaying the bad metrics for a while, while others like the one below just started a few weeks ago. End user first noticed this today. My current thought is that this is a change brought on by a recent patch on the RHEL node itself as it showed up on this particular node just after a reboot and patching. Waiting to see what support comes back with.

aLTeReGo

I believe this has been discussed and explained in great depth in the following thread. Re: Linux Memory Utilization Monitors and You

The TLDR is that NET-SNMP provides this information including things into the value that not everyone agrees should. Unfortunately, they also don't break those other sources of memory usage elsewhere, so they can't be subtracted from the whole. The best solution for accurate memory monitoring on Linux hosts is to utilize the Orion Agent.

stevenstadel

Heard back from support that it is a known issue and development is working on a fix.

pratikmehta003

i have started seeing this for 4 servers and as per support it was found out that the OID which Orion polls is getting negative value from device end.

We are also getting this error when running snmpwalk:

Working with our technical team to see if they can help and if there is any issue in the configuration at device level.

aLTeReGo

Perhaps this is an issue with the version of NET-SNMP being used. What version are you currently running? Have you tried updating your version of NET-SNMP?

pratikmehta003

i have requested my Linux team to check on it... will update once i have the information from them