Jump to content

Оптимизация сетевой подсистемы

Megas
 Share

Recommended Posts

Megas

Megas

Posted
Posted

Сервер.

# cat /proc/cpuinfo

processor : 0

vendor_id : GenuineIntel

cpu family : 6

model : 23

model name : Intel® Celeron® CPU E3400 @ 2.60GHz

stepping : 10

cpu MHz : 2621.586

cache size : 1024 KB

physical id : 0

siblings : 2

core id : 0

cpu cores : 2

apicid : 0

fdiv_bug : no

hlt_bug : no

f00f_bug : no

coma_bug : no

fpu : yes

fpu_exception : yes

cpuid level : 13

wp : yes

flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm constant_tsc pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr lahf_lm

bogomips : 5243.17

 

processor : 1

vendor_id : GenuineIntel

cpu family : 6

model : 23

model name : Intel® Celeron® CPU E3400 @ 2.60GHz

stepping : 10

cpu MHz : 2621.586

cache size : 1024 KB

physical id : 0

siblings : 2

core id : 1

cpu cores : 2

apicid : 1

fdiv_bug : no

hlt_bug : no

f00f_bug : no

coma_bug : no

fpu : yes

fpu_exception : yes

cpuid level : 13

wp : yes

flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm constant_tsc pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr lahf_lm

bogomips : 5242.91

#

 

С картами

 

02:00.0 Ethernet controller: Intel Corporation 82541PI Gigabit Ethernet Controller (rev 05)

Subsystem: Intel Corporation PRO/1000 GT Desktop Adapter

Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 209

Memory at febe0000 (32-bit, non-prefetchable)

Memory at febc0000 (32-bit, non-prefetchable)

I/O ports at ec00

Expansion ROM at 50000000 [disabled]

Capabilities: [dc] Power Management version 2

Capabilities: [e4] PCI-X non-bridge device

Kernel driver in use: e1000

Kernel modules: e1000

 

02:01.0 Ethernet controller: Intel Corporation 82541PI Gigabit Ethernet Controller (rev 05)

Subsystem: Intel Corporation PRO/1000 GT Desktop Adapter

Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 169

Memory at feb80000 (32-bit, non-prefetchable)

Memory at feb60000 (32-bit, non-prefetchable)

I/O ports at e880

Expansion ROM at 50020000 [disabled]

Capabilities: [dc] Power Management version 2

Capabilities: [e4] PCI-X non-bridge device

Kernel driver in use: e1000

Kernel modules: e1000

#

 

ip 192.168.10.254 и алиас 192.168.230.1

 

# ethtool -S eth1

NIC statistics:

rx_packets: 116327680

tx_packets: 175491901

rx_bytes: 20380783521

tx_bytes: 216747380072

rx_broadcast: 95875

tx_broadcast: 1406182

rx_multicast: 136998

tx_multicast: 30

rx_errors: 0

tx_errors: 0

tx_dropped: 0

multicast: 136998

collisions: 0

rx_length_errors: 0

rx_over_errors: 0

rx_crc_errors: 0

rx_frame_errors: 0

rx_no_buffer_count: 0

rx_missed_errors: 0

tx_aborted_errors: 0

tx_carrier_errors: 0

tx_fifo_errors: 0

tx_heartbeat_errors: 0

tx_window_errors: 0

tx_abort_late_coll: 0

tx_deferred_ok: 0

tx_single_coll_ok: 0

tx_multi_coll_ok: 0

tx_timeout_count: 0

tx_restart_queue: 402

rx_long_length_errors: 0

rx_short_length_errors: 0

rx_align_errors: 0

tx_tcp_seg_good: 0

tx_tcp_seg_failed: 0

rx_flow_control_xon: 0

rx_flow_control_xoff: 0

tx_flow_control_xon: 0

tx_flow_control_xoff: 0

rx_long_byte_count: 20380783521

rx_csum_offload_good: 111696431

rx_csum_offload_errors: 61

rx_header_split: 0

alloc_rx_buff_failed: 0

tx_smbus: 0

rx_smbus: 0

dropped_smbus: 0

#

 

Выглядит как:

 

GW_PROV --- SERVER --- SWITH --- BRIDGE

+--- M5

 

ping на бридж и m5

64 bytes from 192.168.230.12: icmp_req=11725 ttl=64 time=1.17 ms

64 bytes from 192.168.230.12: icmp_req=11726 ttl=64 time=0.931 ms

64 bytes from 192.168.230.12: icmp_req=11727 ttl=64 time=1.14 ms

64 bytes from 192.168.230.12: icmp_req=11728 ttl=64 time=1.80 ms

64 bytes from 192.168.230.12: icmp_req=11729 ttl=64 time=0.908 ms

64 bytes from 192.168.230.12: icmp_req=11730 ttl=64 time=2.25 ms

64 bytes from 192.168.230.12: icmp_req=11731 ttl=64 time=1.69 ms

64 bytes from 192.168.230.12: icmp_req=11732 ttl=64 time=1.68 ms

64 bytes from 192.168.230.12: icmp_req=11733 ttl=64 time=1.29 ms

64 bytes from 192.168.230.12: icmp_req=11734 ttl=64 time=2.32 ms

64 bytes from 192.168.230.12: icmp_req=11735 ttl=64 time=1.25 ms

64 bytes from 192.168.230.12: icmp_req=11736 ttl=64 time=1.60 ms

64 bytes from 192.168.230.12: icmp_req=11737 ttl=64 time=2.19 ms

Другими словами все отлично.

 

А вот пинг на интерфейс сервера выгладит не много удручающе.

1508 bytes from 192.168.10.254: icmp_req=12005 ttl=64 time=10.8 ms

1508 bytes from 192.168.10.254: icmp_req=12006 ttl=64 time=9.69 ms

1508 bytes from 192.168.10.254: icmp_req=12007 ttl=64 time=18.5 ms

1508 bytes from 192.168.10.254: icmp_req=12008 ttl=64 time=8.17 ms

1508 bytes from 192.168.10.254: icmp_req=12009 ttl=64 time=19.2 ms

1508 bytes from 192.168.10.254: icmp_req=12011 ttl=64 time=13.7 ms

1508 bytes from 192.168.10.254: icmp_req=12012 ttl=64 time=16.5 ms

1508 bytes from 192.168.10.254: icmp_req=12013 ttl=64 time=13.6 ms

1508 bytes from 192.168.10.254: icmp_req=12014 ttl=64 time=35.3 ms

1508 bytes from 192.168.10.254: icmp_req=12015 ttl=64 time=10.7 ms

1508 bytes from 192.168.10.254: icmp_req=12016 ttl=64 time=11.6 ms

1508 bytes from 192.168.10.254: icmp_req=12017 ttl=64 time=123 ms

1508 bytes from 192.168.10.254: icmp_req=12018 ttl=64 time=18.2 ms

1508 bytes from 192.168.10.254: icmp_req=12019 ttl=64 time=12.7 ms

1508 bytes from 192.168.10.254: icmp_req=12020 ttl=64 time=21.7 ms

1508 bytes from 192.168.10.254: icmp_req=12021 ttl=64 time=37.6 ms

1508 bytes from 192.168.10.254: icmp_req=12022 ttl=64 time=13.3 ms

1508 bytes from 192.168.10.254: icmp_req=12023 ttl=64 time=27.7 ms

1508 bytes from 192.168.10.254: icmp_req=12024 ttl=64 time=99.0 ms

1508 bytes from 192.168.10.254: icmp_req=12027 ttl=64 time=87.2 ms

1508 bytes from 192.168.10.254: icmp_req=12028 ttl=64 time=10.8 ms

1508 bytes from 192.168.10.254: icmp_req=12029 ttl=64 time=66.9 ms

1508 bytes from 192.168.10.254: icmp_req=12030 ttl=64 time=133 ms

1508 bytes from 192.168.10.254: icmp_req=12031 ttl=64 time=14.6 ms

1508 bytes from 192.168.10.254: icmp_req=12032 ttl=64 time=12.7 ms

1508 bytes from 192.168.10.254: icmp_req=12033 ttl=64 time=17.6 ms

1508 bytes from 192.168.10.254: icmp_req=12034 ttl=64 time=14.8 ms

1508 bytes from 192.168.10.254: icmp_req=12035 ttl=64 time=19.3 ms

1508 bytes from 192.168.10.254: icmp_req=12036 ttl=64 time=69.7 ms

1508 bytes from 192.168.10.254: icmp_req=12037 ttl=64 time=13.4 ms

1508 bytes from 192.168.10.254: icmp_req=12038 ttl=64 time=11.0 ms

1508 bytes from 192.168.10.254: icmp_req=12039 ttl=64 time=13.2 ms

1508 bytes from 192.168.10.254: icmp_req=12040 ttl=64 time=11.2 ms

1508 bytes from 192.168.10.254: icmp_req=12041 ttl=64 time=15.6 ms

1508 bytes from 192.168.10.254: icmp_req=12042 ttl=64 time=32.5 ms

1508 bytes from 192.168.10.254: icmp_req=12043 ttl=64 time=16.2 ms

1508 bytes from 192.168.10.254: icmp_req=12044 ttl=64 time=14.8 ms

1508 bytes from 192.168.10.254: icmp_req=12045 ttl=64 time=17.1 ms

1508 bytes from 192.168.10.254: icmp_req=12046 ttl=64 time=65.1 ms

1508 bytes from 192.168.10.254: icmp_req=12047 ttl=64 time=16.4 ms

1508 bytes from 192.168.10.254: icmp_req=12048 ttl=64 time=15.5 ms

1508 bytes from 192.168.10.254: icmp_req=12049 ttl=64 time=16.9 ms

Появляются кучи непонятных задержек в пакетах, что влияет на общую производительность.

 

На сервере работает

Linux prov.local 2.6.18-238.12.1.el5.028stab091.1 #1 SMP Wed Jun 1 13:55:18 MSD 2011 i686 i686 i386 GNU/Linux

 

# cat /etc/redhat-release

CentOS release 5.6 (Final)

#

 

 

 

Сделал:

 

# cat /proc/interrupts

CPU0 CPU1

0: 174332251 30145 IO-APIC-edge timer

1: 2 6 IO-APIC-edge i8042

7: 1 1 IO-APIC-edge parport0

8: 2 1 IO-APIC-edge rtc

9: 0 0 IO-APIC-level acpi

12: 2 2 IO-APIC-edge i8042

169: 15694932 122872645 IO-APIC-level eth1

201: 2206 6975472 IO-APIC-level ata_piix

209: 1484 163388038 IO-APIC-level eth0

NMI: 0 0

LOC: 175393403 175393402

RES: 12289823 983840

ERR: 0

MIS: 0

#

 

 

Не много помогло, но все равно ерунда полная.

Куда еще можно посмотреть и на что глянуть?

 

 

Скоро надо разворачивать еще ppptp и pppoe, но эта машинка явно не потянет это дело.

Megas

Megas

Posted
  • Author
Posted
# ethtool -k eth1

Offload parameters for eth1:

Cannot get device udp large send offload settings: Operation not supported

rx-checksumming: on

tx-checksumming: on

scatter-gather: on

tcp segmentation offload: on

udp fragmentation offload: off

generic segmentation offload: off

generic-receive-offload: off

#

 

# ethtool -g eth1

Ring parameters for eth1:

Pre-set maximums:

RX: 4096

RX Mini: 0

RX Jumbo: 0

TX: 4096

Current hardware settings:

RX: 256

RX Mini: 0

RX Jumbo: 0

TX: 256

#

 

 

 

 

# modinfo igb

filename: /lib/modules/2.6.18-238.12.1.el5.028stab091.1/kernel/drivers/net/igb/igb.ko

version: 2.1.0-k2-1

license: GPL

description: Intel® Gigabit Ethernet Network Driver

author: Intel Corporation, <e1000-devel@lists.sourceforge.net>

srcversion: 53DE31944CEBA8B3896B7E8

alias: pci:v00008086d000010D6sv*sd*bc*sc*i*

alias: pci:v00008086d000010A9sv*sd*bc*sc*i*

alias: pci:v00008086d000010A7sv*sd*bc*sc*i*

alias: pci:v00008086d000010E8sv*sd*bc*sc*i*

alias: pci:v00008086d00001526sv*sd*bc*sc*i*

alias: pci:v00008086d0000150Dsv*sd*bc*sc*i*

alias: pci:v00008086d000010E7sv*sd*bc*sc*i*

alias: pci:v00008086d000010E6sv*sd*bc*sc*i*

alias: pci:v00008086d00001518sv*sd*bc*sc*i*

alias: pci:v00008086d0000150Asv*sd*bc*sc*i*

alias: pci:v00008086d000010C9sv*sd*bc*sc*i*

alias: pci:v00008086d00001516sv*sd*bc*sc*i*

alias: pci:v00008086d00001511sv*sd*bc*sc*i*

alias: pci:v00008086d00001510sv*sd*bc*sc*i*

alias: pci:v00008086d0000150Fsv*sd*bc*sc*i*

alias: pci:v00008086d0000150Esv*sd*bc*sc*i*

alias: pci:v00008086d00001524sv*sd*bc*sc*i*

alias: pci:v00008086d00001523sv*sd*bc*sc*i*

alias: pci:v00008086d00001522sv*sd*bc*sc*i*

alias: pci:v00008086d00001521sv*sd*bc*sc*i*

depends: dca,8021q

vermagic: 2.6.18-238.12.1.el5.028stab091.1 SMP mod_unload 686 REGPARM gcc-4.1

parm: max_vfs:Maximum number of virtual functions to allocate per physical function (uint)

#

 

Собственно вот все что есть.

Megas

Megas

Posted
  • Author
Posted

старое:

# modinfo e1000

filename: /lib/modules/2.6.18-238.12.1.el5.028stab091.1/kernel/drivers/net/e1000/e1000.ko

version: 7.3.21-k4-3-NAPI

license: GPL

description: Intel® PRO/1000 Network Driver

author: Intel Corporation, <linux.nics@intel.com>

srcversion: 2B6567746A6C60FF5C50F4D

alias: pci:v00008086d000010B5sv*sd*bc*sc*i*

alias: pci:v00008086d00001099sv*sd*bc*sc*i*

alias: pci:v00008086d0000108Asv*sd*bc*sc*i*

alias: pci:v00008086d0000107Csv*sd*bc*sc*i*

alias: pci:v00008086d0000107Bsv*sd*bc*sc*i*

alias: pci:v00008086d0000107Asv*sd*bc*sc*i*

alias: pci:v00008086d00001079sv*sd*bc*sc*i*

alias: pci:v00008086d00001078sv*sd*bc*sc*i*

alias: pci:v00008086d00001077sv*sd*bc*sc*i*

alias: pci:v00008086d00001076sv*sd*bc*sc*i*

alias: pci:v00008086d00001075sv*sd*bc*sc*i*

alias: pci:v00008086d00001028sv*sd*bc*sc*i*

alias: pci:v00008086d00001027sv*sd*bc*sc*i*

alias: pci:v00008086d00001026sv*sd*bc*sc*i*

alias: pci:v00008086d0000101Esv*sd*bc*sc*i*

alias: pci:v00008086d0000101Dsv*sd*bc*sc*i*

alias: pci:v00008086d0000101Asv*sd*bc*sc*i*

alias: pci:v00008086d00001019sv*sd*bc*sc*i*

alias: pci:v00008086d00001018sv*sd*bc*sc*i*

alias: pci:v00008086d00001017sv*sd*bc*sc*i*

alias: pci:v00008086d00001016sv*sd*bc*sc*i*

alias: pci:v00008086d00001015sv*sd*bc*sc*i*

alias: pci:v00008086d00001014sv*sd*bc*sc*i*

alias: pci:v00008086d00001013sv*sd*bc*sc*i*

alias: pci:v00008086d00001012sv*sd*bc*sc*i*

alias: pci:v00008086d00001011sv*sd*bc*sc*i*

alias: pci:v00008086d00001010sv*sd*bc*sc*i*

alias: pci:v00008086d0000100Fsv*sd*bc*sc*i*

alias: pci:v00008086d0000100Esv*sd*bc*sc*i*

alias: pci:v00008086d0000100Dsv*sd*bc*sc*i*

alias: pci:v00008086d0000100Csv*sd*bc*sc*i*

alias: pci:v00008086d00001009sv*sd*bc*sc*i*

alias: pci:v00008086d00001008sv*sd*bc*sc*i*

alias: pci:v00008086d00001004sv*sd*bc*sc*i*

alias: pci:v00008086d00001001sv*sd*bc*sc*i*

alias: pci:v00008086d00001000sv*sd*bc*sc*i*

depends: e1000e

vermagic: 2.6.18-238.12.1.el5.028stab091.1 SMP mod_unload 686 REGPARM gcc-4.1

parm: TxDescriptors:Number of transmit descriptors (array of int)

parm: TxDescPower:Binary exponential size (2^X) of each transmit descriptor (array of int)

parm: RxDescriptors:Number of receive descriptors (array of int)

parm: Speed:Speed setting (array of int)

parm: Duplex:Duplex setting (array of int)

parm: AutoNeg:Advertised auto-negotiation setting (array of int)

parm: FlowControl:Flow Control setting (array of int)

parm: XsumRX:Disable or enable Receive Checksum offload (array of int)

parm: TxIntDelay:Transmit Interrupt Delay (array of int)

parm: TxAbsIntDelay:Transmit Absolute Interrupt Delay (array of int)

parm: RxIntDelay:Receive Interrupt Delay (array of int)

parm: RxAbsIntDelay:Receive Absolute Interrupt Delay (array of int)

parm: InterruptThrottleRate:Interrupt Throttling Rate (array of int)

parm: SmartPowerDownEnable:Enable PHY smart power down (array of int)

parm: KumeranLockLoss:Enable Kumeran lock loss workaround (array of int)

parm: copybreak:Maximum size of packet that is copied to a new buffer on receive (uint)

parm: debug:Debug level (0=none,...,16=all) (int)

#

 

 

новое:

 

# modinfo e1000

filename: /lib/modules/2.6.18-238.19.1.el5/kernel/drivers/net/e1000/e1000.ko

version: 8.0.30-NAPI

license: GPL

description: Intel® PRO/1000 Network Driver

author: Intel Corporation, <linux.nics@intel.com>

srcversion: 1D436A0D02B18BE145D2AE1

alias: pci:v00008086d000010B5sv*sd*bc*sc*i*

alias: pci:v00008086d00001099sv*sd*bc*sc*i*

alias: pci:v00008086d0000108Asv*sd*bc*sc*i*

alias: pci:v00008086d0000107Csv*sd*bc*sc*i*

alias: pci:v00008086d0000107Bsv*sd*bc*sc*i*

alias: pci:v00008086d0000107Asv*sd*bc*sc*i*

alias: pci:v00008086d00001079sv*sd*bc*sc*i*

alias: pci:v00008086d00001078sv*sd*bc*sc*i*

alias: pci:v00008086d00001077sv*sd*bc*sc*i*

alias: pci:v00008086d00001076sv*sd*bc*sc*i*

alias: pci:v00008086d00001075sv*sd*bc*sc*i*

alias: pci:v00008086d00001028sv*sd*bc*sc*i*

alias: pci:v00008086d00001027sv*sd*bc*sc*i*

alias: pci:v00008086d00001026sv*sd*bc*sc*i*

alias: pci:v00008086d0000101Esv*sd*bc*sc*i*

alias: pci:v00008086d0000101Dsv*sd*bc*sc*i*

alias: pci:v00008086d0000101Asv*sd*bc*sc*i*

alias: pci:v00008086d00001019sv*sd*bc*sc*i*

alias: pci:v00008086d00001018sv*sd*bc*sc*i*

alias: pci:v00008086d00001017sv*sd*bc*sc*i*

alias: pci:v00008086d00001016sv*sd*bc*sc*i*

alias: pci:v00008086d00001015sv*sd*bc*sc*i*

alias: pci:v00008086d00001014sv*sd*bc*sc*i*

alias: pci:v00008086d00001013sv*sd*bc*sc*i*

alias: pci:v00008086d00001012sv*sd*bc*sc*i*

alias: pci:v00008086d00001011sv*sd*bc*sc*i*

alias: pci:v00008086d00001010sv*sd*bc*sc*i*

alias: pci:v00008086d0000100Fsv*sd*bc*sc*i*

alias: pci:v00008086d0000100Esv*sd*bc*sc*i*

alias: pci:v00008086d0000100Dsv*sd*bc*sc*i*

alias: pci:v00008086d0000100Csv*sd*bc*sc*i*

alias: pci:v00008086d00001009sv*sd*bc*sc*i*

alias: pci:v00008086d00001008sv*sd*bc*sc*i*

alias: pci:v00008086d00001004sv*sd*bc*sc*i*

alias: pci:v00008086d00001001sv*sd*bc*sc*i*

alias: pci:v00008086d00001000sv*sd*bc*sc*i*

depends: 8021q

vermagic: 2.6.18-238.19.1.el5 SMP mod_unload 686 REGPARM 4KSTACKS gcc-4.1

parm: TxDescriptors:Number of transmit descriptors (array of int)

parm: TxDescPower:Binary exponential size (2^X) of each transmit descriptor (array of int)

parm: RxDescriptors:Number of receive descriptors (array of int)

parm: TxDescriptorStep:Transmit Descriptor Step (array of int)

parm: Speed:Speed setting (array of int)

parm: Duplex:Duplex setting (array of int)

parm: AutoNeg:Advertised auto-negotiation setting (array of int)

parm: FlowControl:Flow Control setting (array of int)

parm: XsumRX:Disable or enable Receive Checksum offload (array of int)

parm: TxIntDelay:Transmit Interrupt Delay (array of int)

parm: TxAbsIntDelay:Transmit Absolute Interrupt Delay (array of int)

parm: RxIntDelay:Receive Interrupt Delay (array of int)

parm: RxAbsIntDelay:Receive Absolute Interrupt Delay (array of int)

parm: InterruptThrottleRate:Interrupt Throttling Rate (array of int)

parm: SmartPowerDownEnable:Enable PHY smart power down (array of int)

parm: copybreak:Maximum size of packet that is copied to a new buffer on receive (uint)

parm: ignore_64bit_dma:Ignore 64-bit DMA (DAC) capability (int)

parm: debug:Debug level (0=none,...,16=all) (int)

#

 

 

Сейчас правда загрузки всего 11-15мегабит и порядка 5к pps, но пока вот такие результаты:

1508 bytes from 192.168.10.254: icmp_req=327 ttl=64 time=6.82 ms

1508 bytes from 192.168.10.254: icmp_req=328 ttl=64 time=16.7 ms

1508 bytes from 192.168.10.254: icmp_req=329 ttl=64 time=8.46 ms

1508 bytes from 192.168.10.254: icmp_req=330 ttl=64 time=12.1 ms

1508 bytes from 192.168.10.254: icmp_req=331 ttl=64 time=8.40 ms

1508 bytes from 192.168.10.254: icmp_req=332 ttl=64 time=18.2 ms

1508 bytes from 192.168.10.254: icmp_req=333 ttl=64 time=6.65 ms

1508 bytes from 192.168.10.254: icmp_req=334 ttl=64 time=9.34 ms

1508 bytes from 192.168.10.254: icmp_req=335 ttl=64 time=18.9 ms

1508 bytes from 192.168.10.254: icmp_req=336 ttl=64 time=12.5 ms

1508 bytes from 192.168.10.254: icmp_req=337 ttl=64 time=11.5 ms

1508 bytes from 192.168.10.254: icmp_req=338 ttl=64 time=14.3 ms

1508 bytes from 192.168.10.254: icmp_req=339 ttl=64 time=7.25 ms

1508 bytes from 192.168.10.254: icmp_req=340 ttl=64 time=8.55 ms

1508 bytes from 192.168.10.254: icmp_req=341 ttl=64 time=9.07 ms

1508 bytes from 192.168.10.254: icmp_req=342 ttl=64 time=9.94 ms

1508 bytes from 192.168.10.254: icmp_req=343 ttl=64 time=11.6 ms

1508 bytes from 192.168.10.254: icmp_req=344 ttl=64 time=18.4 ms

1508 bytes from 192.168.10.254: icmp_req=345 ttl=64 time=15.9 ms

1508 bytes from 192.168.10.254: icmp_req=346 ttl=64 time=8.83 ms

1508 bytes from 192.168.10.254: icmp_req=347 ttl=64 time=8.11 ms

1508 bytes from 192.168.10.254: icmp_req=348 ttl=64 time=11.0 ms

1508 bytes from 192.168.10.254: icmp_req=349 ttl=64 time=10.7 ms

1508 bytes from 192.168.10.254: icmp_req=350 ttl=64 time=8.24 ms

1508 bytes from 192.168.10.254: icmp_req=351 ttl=64 time=10.7 ms

1508 bytes from 192.168.10.254: icmp_req=352 ttl=64 time=12.2 ms

1508 bytes from 192.168.10.254: icmp_req=353 ttl=64 time=11.5 ms

 

Вроде как частично в чем-то помогло, нету заржек более 100ms, но как говориться еще не вечер, скоро нагрузка будет подниматся и тогда будет видно

Abram

Abram

Posted
Posted

Но покажите-ка еще ... и modinfo igb.

Да при чем тут igb, если там древнючие сетевухи стоят?

 

2 Megas - попробуйте для начала собрать последние драйвера для этих адаптеров.

Пардон, не заметил. Я думал, таких уже не бывает.

Megas,

И да - попробуйте драйвера с сайта Интела.

Еще попробуйте увеличить ring buffer до 4096 (это где-то в районе ethtool -g).

Megas

Megas

Posted
  • Author
Posted
ethtool -G eth1 rx 4096 tx 4096

 

# ethtool -g eth1

Ring parameters for eth1:

Pre-set maximums:

RX: 4096

RX Mini: 0

RX Jumbo: 0

TX: 4096

Current hardware settings:

RX: 4096

RX Mini: 0

RX Jumbo: 0

TX: 4096

#

 

 

1508 bytes from 192.168.10.254: icmp_req=3339 ttl=64 time=25.5 ms

1508 bytes from 192.168.10.254: icmp_req=3340 ttl=64 time=79.4 ms

1508 bytes from 192.168.10.254: icmp_req=3341 ttl=64 time=10.9 ms

1508 bytes from 192.168.10.254: icmp_req=3342 ttl=64 time=12.7 ms

1508 bytes from 192.168.10.254: icmp_req=3343 ttl=64 time=13.1 ms

1508 bytes from 192.168.10.254: icmp_req=3344 ttl=64 time=114 ms

1508 bytes from 192.168.10.254: icmp_req=3346 ttl=64 time=56.1 ms

1508 bytes from 192.168.10.254: icmp_req=3347 ttl=64 time=8.27 ms

1508 bytes from 192.168.10.254: icmp_req=3348 ttl=64 time=115 ms

1508 bytes from 192.168.10.254: icmp_req=3349 ttl=64 time=83.3 ms

1508 bytes from 192.168.10.254: icmp_req=3350 ttl=64 time=66.4 ms

1508 bytes from 192.168.10.254: icmp_req=3351 ttl=64 time=12.2 ms

1508 bytes from 192.168.10.254: icmp_req=3352 ttl=64 time=10.8 ms

1508 bytes from 192.168.10.254: icmp_req=3353 ttl=64 time=12.0 ms

1508 bytes from 192.168.10.254: icmp_req=3354 ttl=64 time=94.7 ms

 

один черт, на соседнее устройство рядом с этим сервом пинг 2-3мс, а на сервер прыгает что подорванный.

 

Дрова обовил, количество буферов поменял.

Megas

Megas

Posted
  • Author
Posted

martini, спасибо, уже увидел, лажанулся, но зато сейчас смотрю что при пике в 30мегабит как раз 1508 байт проходят на ура.

  • 2 months later...
Megas

Megas

Posted
  • Author
Posted

В общем снова полезли какие-то траблы, только вот теперь ширина канала по более, порядка 200мегабит.

А вот выжать максимум что-то не получается.

И как правильно диагностировать производительность канала тоже не совсем понятно.

К стати, чтиво:

http://www.redbooks.ibm.com/redpapers/pdfs/redp4285.pdf

кто нибудь осиливал?

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
 Share
Go to topic listing
×
×
  • Create New...
На сайте используются файлы cookie и сервисы аналитики для корректной работы форума и улучшения качества обслуживания. Продолжая использовать сайт, вы соглашаетесь с использованием файлов cookie и с Политикой конфиденциальности.