packet loss

romeor · #1 August 7th, 2012, 14:45

Hi,

Got FreeBSD 9.0 with PF and hfsc.

Code:

9.0-RELEASE-p3

Code:

QUEUE                             BW SCH  PRIO     PKTS    BYTES   DROP_P   DROP_B QLEN BORROW SUSPEN     P/S     B/S
root_em0                        199M hfsc    0        0        0        0        0    0                     0       0
 std_up                         111M hfsc      17394182   24976M  4189660 6161010K    0                  9589  14429K
 ftp_up                          43M hfsc       9082518   12420M      392   562128   17                  5531 7931782
 sig_up                          43M hfsc            36     6723        0        0    0                     0       0
root_em1                        199M hfsc    0        0        0        0        0    0                     0       0
 std_down                       111M hfsc        480341 63870408        0        0    0                   261   36663
 ftp_down                        43M hfsc       3928034  253654K        0        0    0                  2338  154329
 sig_down                        43M hfsc           216   146993        0        0    0                     0       0

In such situation I've got this:

Code:

62 packets transmitted, 54 packets received, 12.9% packet loss
round-trip min/avg/max/stddev = 3.795/12.217/19.819/5.129 ms

some debugging:

Code:

last pid:   401;  load averages:  0.33,  0.34,  0.25                                                                                                             up 12+19:42:15  16:43:00
26 processes:  1 running, 25 sleeping
CPU:  0.0% user,  0.0% nice,  7.5% system,  0.0% interrupt, 92.5% idle
Mem: 49M Active, 1132M Inact, 1212M Wired, 412K Cache, 826M Buf, 5502M Free
Swap: 9267M Total, 9267M Free

Code:

    1 users    Load  0.37  0.35  0.26                  Aug  7 16:43

Mem:KB    REAL            VIRTUAL                       VN PAGER   SWAP PAGER
        Tot   Share      Tot    Share    Free           in   out     in   out
Act   62112    6900   585068     7544 5633064  count
All  121460    7312 1074418k    16532          pages
Proc:                                                            Interrupts
  r   p   d   s   w   Csw  Trp  Sys  Int  Sof  Flt        cow   11804 total
             31       23k    3  211  11k 1025             zfod      4 atapci0+
                                                          ozfod       ehci0 21
 5.4%Sys   0.0%Intr  0.0%User  0.0%Nice 94.6%Idle        %ozfod       ehci1 23
|    |    |    |    |    |    |    |    |    |    |       daefr  6843 em0 256
===                                                       prcfr  3611 em1 257
                                        14 dtbuf          totfr  1132 hpet0:t0
Namei     Name-cache   Dir-cache    206234 desvn          react   180 hpet0:t1
   Calls    hits   %    hits   %    181213 numvn          pdwak    28 hpet0:t2
       3       3 100                 51551 frevn          pdpgs     6 hpet0:t3
                                                          intrn
Disks  ada0  ada1 pass0 pass1                     1241492 wire
KB/t  16.25 16.25  0.00  0.00                       50828 act
tps       2     2     0     0                     1159560 inact
MB/s   0.03  0.03  0.00  0.00                         412 cache
%busy     0     0     0     0                     5632652 free
                                                   846240 buf

Can anyone give any advice on this topic? I've controlled the L2 from the PoP to server - seems fine, no errors nor CRC.
The main problem is that OSPF and BGP sessions go down because of time-outs (using Quagga on the same machine)

romeor · #2 August 8th, 2012, 15:41

I also noticed, that I've got some Oerrs in netstat -i stats for my dot1q VLANs, but no errors on parent interfaces and there are 0 Colls and Oerrs are about 8% of Opkts. A this moment I configured switch port and interface cards to be in full-duplex 1000baseT mode manually (for any case). Now just watching it. My NICs are not built-in.

As there are no Colls, I may assume, that these errors come from ALTQ rules? any ideas?

update:
Yes, as soon as I try to send more traffic, than allowed by ALTQ pipe, there are errors in netstat -i . Is it normal ?

romeor · #3 August 9th, 2012, 10:53

During high UDP traffic ( 14k pps, 160 Mbps), got this:

Code:

            input     (vlan1025)           output
   packets  errs idrops      bytes    packets  errs      bytes colls
        74     0     0       9226      14572   358   22481884     0
        73     0     0       7295      14361   160   21858836     0
        66     0     0       6076      13998    62   21173122     0
        59     0     0       5288      15105     0   22743880     0
        82     0     0      11147      14639   597   22940533     0
       101     0     0      41152      14219   533   22192901     0
        60     0     0       5424      13863     0   20874934     0
        70     0     0       8704      14344   447   22272502     0

Is it normal ?

just curious.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Solved] KVM virtual server FreeBSD guest packet loss	AndyUKG	General	1	June 13th, 2012 16:55
[PF] [NAT and ROUTING] Packet loss on FreeBSD 9.0	pl0p	Firewalls	4	May 25th, 2012 17:31
IPv6 + PF + packet loss	folivora	Networking	4	February 24th, 2012 16:38
packet loss in IPFW	atwinix	Firewalls	0	June 24th, 2011 10:36
Packet loss...	laneo	Networking	3	November 6th, 2009 02:27