Quantcast
Channel: THWACK: All Content - All Communities
Viewing all articles
Browse latest Browse all 20518

Resource: Top 10/25/XX Errors (only)

$
0
0

Solarwinds Top 10/25/xx Errors and Discards does not provide usable information in any sizable or diverse environment.

 

We need a resource for

 

      TOP 10/25/X ERRORS

 

Each SW company I've worked in found the existing resource to be useless due to the presence of ASAs alone.  Other diverse situations fill the Top 10/25 with noisy discards--that are NORMAL. Errors never show up because they are drowned out by normal, unavoidable discards.

 

Errors are ALWAYS BAD in volume and MUST be eiminated.

 

Discards are a normal occurrence in many situations. They are not an "issue", however our top 25 is filled with discards.  Meanwhile we have servers around the country with 100,000s of physical errors that we don't know about.  Cisco ASAs report every dropped/denied packet as a ethernet discard, something I believe is utter nonsense. This was a denial by the firewall "application", not by ethernet NIC. If my Windows OS ignores multicast packets because they aren't for it, it isn't an Ethernet discard.  If flow control between the NIC and switch are working, the NIC can back the switch off from flooding heavy traffic, causing discards on the switch port buffers.  While this is a sign the device may be overloaded, it isn't a "network" issue but a server/app capacity issue.  We even have a Cisco 3750 that randomly reports counter rollovers as 4 billion discards.  We've tried many things to fix it, but everyday we have dozens of interfaces with 4,8,12 billion discards.  Even after eliminating them (by NOT monitoring at all), we're left with useless ASA discards.

 

Meanwhile, we're seeing massive discards on our ESX trunks, we think because the switches weren't configured with portfast trunk.  The Lab ESX, rebooted constantly has far more discards than all the production ESXes combined, yet does the least and has the lowest traffic.  We think this is due to STP blocking upon resets.  This is a rare instance where a discard may be avoided.   Discards can occur with duplex/speed mismatch, but they usually occur in combination with other errors.  As it stands now, top XX discards is never helpful, except in a small shop.


Viewing all articles
Browse latest Browse all 20518

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>