2.3.2 Upgrade - internet is suddenly super slow
-
Hi everyone,
Hope I can get some help here. I upgraded to 2.3.2 a few days ago and the internet speeds have dropped tremendously.
I'm on a 100/100 fibre connection but am getting more like 2/8Mbps at the moment. That's on all laptops, desktops, phones, tablets etc. All laptops and desktops are wired so it's not an issue with the AP.
I haven't touched anything else except for upgrading.Here's what I've looked at so far:
The forum… found some posts put none that really seem to have the same problem as myself.
Top looks fine
last pid: 16801; load averages: 0.93, 1.05, 1.13 up 0+00:43:18 22:20:38 61 processes: 1 running, 60 sleeping CPU: 0.6% user, 0.0% nice, 0.0% system, 19.4% interrupt, 80.1% idle Mem: 298M Active, 1073M Inact, 202M Wired, 210M Buf, 378M Free Swap: 4096M Total, 4096M Free PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND 13728 root 1 20 0 21856K 3004K CPU1 1 0:00 0.10% top 30815 root 2 20 0 515M 104M nanslp 3 0:22 0.00% snort 12053 root 2 20 0 515M 106M nanslp 2 0:19 0.00% snort 12643 squid 1 20 0 143M 73276K kqread 1 0:13 0.00% squid 22278 root 1 20 0 16676K 2272K bpf 1 0:03 0.00% filterlog 749 root 1 20 0 14508K 2312K select 2 0:03 0.00% syslogd 31406 root 1 20 0 39136K 7352K kqread 1 0:02 0.00% nginx 31439 root 1 20 0 39136K 7424K kqread 3 0:01 0.00% nginx 49309 root 1 52 20 17000K 2396K wait 2 0:01 0.00% sh [...]Turned off Snort/Squid/Squidguard/Sipproxd which are the only packages I have installed - made no difference.
I'm at a loss as to where to look to try and find out what the problem is.
Any help as to where to look next would be greatly appreciated. -
Did some more digging…
https://doc.pfsense.org/index.php/Low_Throughput_TroubleshootingSo it seems one of the IRQ processes was using around 80% of a CPU core -
12 root -84 - 0K 400K CPU0 0 47:11 79.30% [intr{irq7: ppc0}]
So I followed these instructions:
Another item to check is under System > Advanced on the Networking tab. Ensure that the boxes are checked for Disable hardware TCP segmentation offload and Disable hardware large receive offload. If they are already checked, try toggling Disable hardware checksum offload. If no difference is observed, toggle it back.
Turns out they were already disabled and I unchecked the boxes next to both Disable hardware TCP segmentation offload and Disable hardware large receive offload - rebooted.
Now the high IRQ usage has not gone away but I have full speed again.
12 root -84 - 0K 400K WAIT 0 11:14 74.66% [intr{irq7: ppc0}] -
That's the parallel port driver and should be doing nothing. I believe I remember some previous reports of it spinning on interrupt like that. See if you can disable any parallel ports in your BIOS.
