Latency spikes every 15 minutes after upgrade to 2.6.0 CE
-
After upgrading to 2.6.0 CE, my upstream gateway ping monitoring is picking up significant latency spikes that occur every 15 minutes on the minute. I don't use dpinger for this, just a simple script that pings my upstream router by its IP address and sends the logs to my logger for monitoring and visualization. Here's what that looks like (upgrade was just before 10:30):
I believe a filter reload is triggered every 15 minutes when there are scheduled firewall rules configured, so I removed those (08:10). I only had a few so not a big deal. For others that might be a bigger issue. After that the latency spikes went away and everything seems normal again (which is visible near the tail end of that graph).
So it would seem that filter reloads are not as hitless as they once were in 2.5.2 and earlier? Is this a known issue?
-
E encrypt1d referenced this topic on
-
Same as in 2.5.0 which had people revert to 2.4.5p1 and wait for the 2.5.1 release.
-
I'd say yes...
-
That's the one. Thanks.
-
@encrypt1d said in Latency spikes every 15 minutes after upgrade to 2.6.0 CE:
Is this a known issue
Install the new System Patches package and the "recommended" patch:
Disable pf counter data preservation to temporarily work around latency when reloading large rulesets (Redmine #12827)
