2.0.2-RELEASE (i386) - Status: System logs full of issues
-
Hi Guys, let me ask some questions and describe my network in few words.
I'm based in East Europe -> former Czecho-Slovakia and I'm running network with ~60 computers behind m0n0wall since 2005. m0n0wall is very stable firewall but some features are missing eg. squid, upnp, detailed per user statistic etc. That's the reason I have decided to try pfsense, this is already 2nd time i'm playing with pfSence. ~2years ago after some testing i have decided to stay with m0n0wall due to issues with CaptivePortal and slow "UI response" if using old WRAP board. (during that time i didn't had any other HW)Now, I have build a new PC with Intel(R) Atom(TM) CPU D510 @ 1.66GHz, 4GB RAM, 1 x onboard NIC 1G, and 2nd NIC is some older 100M Realtek chipset, HDD is 500GB 7200RPM for squid caching.
This is fresh installation 2.0.2-RELEASE (i386).
I have installed these packages: 1) squid 2) Lightsquid "reporting" and 3) arpingIn this moment almost everything is left with default settings except of squid, i put these settings: http://doc.pfsense.org/index.php/Squid_Package_Tuning
My pfSense box is running in the same LAN network as my monowall box. LAN 192.168.100.2 (mono is 192.168.100.1). Both boxes are using static Public IP, 213.210.100.7 and 213.210.100.18 for pfSense.I have simply switched 2 PCs within my existing Lan network to use pfSense as gateway.
My issues and Questions:
-
if I do ping 192.168.100.2 (LAN interface); pings are "changing" from 1ms to 200ms. I'm pinging from PC 192.168.100.122, cable connected in a same L2 switch within the rack box
-
This error message is repeating too often:
Jan 24 15:36:25 miniupnpd[51718]: sendto(udp_notify=12, 192.168.100.2): No buffer space available
Jan 24 15:37:25 miniupnpd[51718]: sendto(udp_notify=12, 192.168.100.2): No buffer space available
Jan 24 15:35:59 dnsmasq[29569]: failed to send packet: No buffer space available
Jan 24 15:35:59 dnsmasq[29569]: failed to send packet: No buffer space availableJan 24 14:31:09 kernel: Bump sched buckets to 64 (was 0)
As i mentioned, there are just 2 users connected to this box, which is nothing in terms of performance needs.
Could you please help with some ideas how to get rid of those issues?
Thank you, Peter
-
-
"Bump sched buckets to 64 (was 0) " is normal when using limiters and/or captive portal.
"No buffer space" generally means what it says, the NIC buffer was full (too many packets trying to go out) but it can also happen if the NIC is down or for any reason unable to send out packets.
Do you have captive portal or limiters enabled? If so, exactly what settings are you using?
It would help to see:
cat /tmp/rules.debug ipfw show ipfw pipe show ipfw queue show
It's possible that perhaps you have a limiter configured for a speed so low that it's not able to send packets to your lan properly.
-
2nd NIC is some older 100M Realtek chipset
This could be an issue. 100Mb Realtek NICs have a bad reputation for misbehaving. What chipset/NIC is it?
Steve