Project

General

Profile

Actions

Bug #12919

closed

Enabling gateway failover introduces latency increase and causes artificial failover scenario

Added by Ash Morris about 3 years ago. Updated about 3 years ago.

Status:
Not a Bug
Priority:
Normal
Assignee:
-
Category:
Gateway Monitoring
Target version:
-
Start date:
Due date:
% Done:

0%

Estimated time:
Release Notes:
Default
Affected Plus Version:
Affected Architecture:
All

Description

Forum post: https://forum.netgate.com/topic/170595/sg-5100-wan-failover-at-gigabit-saturation?_=1646751316923

Issue expressed on SG-3100 and SG-5100.

Gateway failover group configured as gigabit interface on igb0 as Tier 1 and 80Mb down interface on igb1 as Tier 2. Successful gateway failover after ~30s is triggered as "Member down" with igb0 RTTsd >1s under heavy load. RTTsd remains <400ms on igb0 under identical heavy load when sert as solitary gateway, interface remains up >5 mins (end of test). Issue is infinitely reproducable.

Actions #1

Updated by Ash Morris about 3 years ago

Forum post: https://forum.netgate.com/topic/170595/sg-5100-wan-failover-at-gigabit-saturation?_=1646751316923

Issue expressed on SG-3100 and SG-5100.

Gateway failover group configured as gigabit DHCP interface on igb0 as Tier 1, 80Mb PPPoE interface on igb1 as Tier 2. Successful gateway failover after ~30s is triggered as "Member down" with igb0 RTTsd >1s when downstream is saturated. RTTsd remains <400ms on igb0 under identical saturation load when set as solitary gateway; interface remains up >5 mins (end of test). Issue is infinitely reproducable.

Actions #2

Updated by Ash Morris about 3 years ago

Apologies, affected version should read 22.01.

Actions #3

Updated by Steve Wheeler about 3 years ago

  • Target version set to 22.05
  • Affected Plus Version changed from 21.02 to 22.01
  • Affected Architecture amd64 added
Actions #4

Updated by Steve Wheeler about 3 years ago

  • Affected Architecture All added
  • Affected Architecture deleted (amd64)
Actions #5

Updated by Ash Morris about 3 years ago

The issue issue is resolved, or rather is not an issue / not an accurate description. The same latency increase to >1s was recorded while testing the solitary gateway config this morning, therefore is no longer confined / attributable to enabling failover.

Actions #6

Updated by Jim Pingle about 3 years ago

  • Status changed from New to Not a Bug
  • Target version deleted (22.05)
  • Affected Plus Version deleted (22.01)

The symptoms sound similar to #12827 -- it might be worth trying the workaround which is available in the recommended patches list in the System Patches package.

If you have any limiters, it may also be worth trying a development snapshot as there is a potentially related fix in the kernel there as well.

Either way, drop a note on your thread with the results if you try those.

Actions

Also available in: Atom PDF