Skip to contentSkip to conversation

speed problem

LARTC Vishal Gandhi Kommineni 5 participants 10 messages Jan 2, 2003 – Jan 4, 2003

lore
Share & tools
lore
#0 lore raw

Hi all,
I have a pecular problem

i have a setup of 3 linux boxes and a windows pc

boxes A, B, C and a pc are connected in the follwoing sequence

A (eth0)< ----> (eth0) B (eth1)< -----> (eth0) C (eth1) <------> PC

1) using ftp I am able to transfer data from B to A both ways, the speeds are around 1.2e
2) I am also able to transfer similarly to and fro from B to PC via box C

but

3) when i am trying to ftp data from A to PC via B and C to and fro the data transfer speed has drastically come down, Please note that in both B and C ipforward is enabled.

Kindly help

Thanks in advance,

Vishal Gandhi

#1 lore raw
Quoted 21 lines · Vishal Gandhi Kommineni

Hi all,
I have a pecular problem

i have a setup of 3 linux boxes and a windows pc

boxes A, B, C and a pc are connected in the follwoing sequence

A (eth0)< ----> (eth0) B (eth1)< -----> (eth0) C (eth1) <------> PC

1) using ftp I am able to transfer data from B to A both ways, the speeds are around 1.2e
2) I am also able to transfer similarly to and fro from B to PC via box C

but

3) when i am trying to ftp data from A to PC via B and C to and fro the data transfer speed has drastically come down,
Please note that in both B and C ipforward is enabled.

Quoted 5 lines

Kindly help

Thanks in advance,

Vishal Gandhi

hi vishal,

the first thing i'd like to check is if the ethernet cards are running in full- or half-duplex mode. AFAIK running one card in half-duplex mode and the other trying to force full-duplex mode can lead to lots of collisions, so this could be a reason for slow speed.

you could use a hub or switch instead of the crossover cable and watch the collision LED (if there is any *g*).

greetings,

ulric

Signature · 6 lines
ulrich schwarz, computing center, university of ulm, germany


_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
#2 lore raw
In reply to Vishal Gandhi Kommineni

Warm wishes for the new year 2003,
thanks for the reply, i found that there are collission when i use a hub in between,
How do i make sure that both the ethernet cards are running in half or full duplex mode,
how do i check if weathere an ethernet is in half or full duplex mode Thanks in advance
Vishal Gandhi

----- Original Message -----
From: "ulrich schwarz" <[email protected]>
To: "Vishal Gandhi Kommineni" <[email protected]>
Cc: "LARTC" <[email protected]>
Sent: Thursday, January 02, 2003 2:34 PM
Subject: Re: [LARTC] speed problem

Quoted 9 lines
Quoted 8 lines · Vishal Gandhi Kommineni

Hi all,
I have a pecular problem

i have a setup of 3 linux boxes and a windows pc

boxes A, B, C and a pc are connected in the follwoing sequence

Quoted 6 lines
Quoted 6 lines

A (eth0)< ----> (eth0) B (eth1)< -----> (eth0) C (eth1) <------> PC

1) using ftp I am able to transfer data from B to A both ways, the speeds are around 1.2e
2) I am also able to transfer similarly to and fro from B to PC via box

C

Quoted 28 lines
Quoted 7 lines

but

3) when i am trying to ftp data from A to PC via B and C to and fro the data transfer speed has drastically come down,
Please note that in both B and C ipforward is enabled.

Quoted 5 lines

Kindly help

Thanks in advance,

Vishal Gandhi

hi vishal,

the first thing i'd like to check is if the ethernet cards are running in full- or half-duplex mode. AFAIK running one card in half-duplex mode and the other trying to force full-duplex mode can lead to lots of collisions, so this could be a reason for slow speed.

you could use a hub or switch instead of the crossover cable and watch the collision LED (if there is any *g*).

greetings,

ulric

Quoted 8 lines

--
ulrich schwarz, computing center, university of ulm, germany

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

#3 lore raw
In reply to Vishal Gandhi Kommineni
Quoted 6 lines · Vishal Gandhi Kommineni

Warm wishes for the new year 2003,
thanks for the reply, i found that there are collission when i use a hub in between,
How do i make sure that both the ethernet cards are running in half or full duplex mode,
how do i check if weathere an ethernet is in half or full duplex mod

Could you provide the output from 'ifconfig' and 'route -n' on A, B and C? This is either a reverse DNS, duplex mismatch or routing problem.

To see if your are full or half duplex, see if ifconfig reports any collisions on the eth interface - on full duplex you never get any collisions.

If it takes a long time from you try to connect until it starts downloading, you have a DNS problem.

If the download is incredibly slow and one or more of the network interfaces on A, B or C report a huge amount of collisions, you have a duplex problem.

Routing problems usually result in not being able to connect at all, which doesn't seem to be the case here.

Signature · 7 lines
The content of this message can either be used according to normal
copyright law (including quoting and personal copies) or by the terms of
the GNU Free Documentation License http://www.gnu.org/licenses/fdl.html

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
#4 lore raw
In reply to Vishal Gandhi Kommineni

Vishal,

: > the first thing i'd like to check is if the ethernet cards are running : > in full- or half-duplex mode. AFAIK running one card in half-duplex mode : > and the other trying to force full-duplex mode can lead to lots of : > collisions, so this could be a reason for slow speed.

I agree with Ulrich, and apparently that may have been your problem.

: thanks for the reply, i found that there are collission when i use a : hub in between, How do i make sure that both the ethernet cards are : running in half or full duplex mode, how do i check if weathere an : ethernet is in half or full duplex mode

You may find that one of the following tools helps you to determine what your ethernet card has negotiated with its peer:

- ethtool ( the tool which will probably continue to be maintained and supported under linux )
http://www.gnu.org/directory/sysadmin/Monitor/ethtool.html

- mii-diag, aka mii-tool, ( one of these two is usually available on common distributions available today )
http://freshmeat.net/projects/mii-diag/

-Martin

Here's some example output from mii-diag and mii-tool:

# mii-tool --help
usage: mii-tool [-VvRrwl] [-A media,... | -F media] [interface ...]
       -V, --version               display version information
       -v, --verbose               more verbose output
       -R, --reset                 reset MII to poweron state
       -r, --restart               restart autonegotiation
       -w, --watch                 monitor for link status changes
       -l, --log                   with -w, write events to syslog
       -A, --advertise=media,...   advertise only specified media
       -F, --force=media           force specified media technology
media: 100baseT4, 100baseTx-FD, 100baseTx-HD, 10baseT-FD, 10baseT-HD,
       (to advertise both HD and FD) 100baseTx, 10baseT
# mii-tool eth0
eth0: negotiated 100baseTx-FD flow-control, link ok
# mii-diag --help
Usage: mii-diag [-aDfrRvVw] [-AF <speed+duplex>] [--watch] <interface>.

Show the link status of the specified network interface. The default interface is "eth0".

The common usage is
mii-diag eth0
Frequently used options are
-A --advertise <speed|setting>
-F --fixed-speed <speed>

        Speed is one of: 100baseT4, 100baseTx, 100baseTx-FD, 100baseTx-HD,
                         10baseT, 10baseT-FD, 10baseT-HD

-s --status Return exit status 2 if there is no link beat.

Less frequently used options are
-a --all-interfaces Show the status all interfaces
(Not recommended with options that change settings.)

   -D  --debug
   -g  --read-parameters        Get driver-specific parameters.
   -G  --set-parameters PARMS   Set driver-specific parameters.
       Parameters are comma separated, missing elements retain existing
value.
   -M  --msg-level LEVEL        Set the driver message bit map.
   -p  --phy ADDR               Set the PHY (MII address) to report.
   -r  --restart        Restart the link autonegotiation.
   -R  --reset          Reset the transceiver.
   -v  --verbose        Report each action taken.
   -V  --version        Emit version information.
   -w  --watch          Continuously monitor the transceiver and report

changes.

This command returns success (zero) if the interface information can be read. If the --status option is passed, a zero return means that the interface has link beat.

# mii-diag eth0
Basic registers of MII PHY #1:  3000 7809 02a8 0154 05e1 0000 0000 0000.
 Basic mode control register 0x3000: Auto-negotiation enabled.
 Basic mode status register 0x7809 ... 7809.
   Link status: not established.
   End of basic transceiver information.
Signature · 5 lines
Martin A. Brown --- SecurePipe, Inc. --- [email protected]

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
#5 lore raw
In reply to Vishal Gandhi Kommineni

I have tried the mii-tool as suggested by [email protected] it did help to some extent but I was not successful in the overall situation I would like to elaborate the setup in more detail

A(eth0) <----->distance 2mts <---> (intel Switch 550T) <---->distance30mts (hub) distance 20mts <--->B(eth0) (eth1 wireless card wan) < ~ ~ ~ ~ ~ > (eth1 wireless card wan) (eth0) <---->distance 2mts (hub) <------>distance 10mts (PC)

A(eth0) the value of this interface card as per the dmesg of the box is as below

eth0: Intel Corporation 82557 [Ethernet Pro 100], 00:90:27:B2:04:82, IRQ 10.
Receiver lock-up bug exists -- enabling work-around.
Board assembly 677173-001, Physical connectors present: RJ45 Primary interface chip i82555 PHY #1.
General self-test: passed.
Serial sub-system self-test: passed.
Internal registers self-test: passed.
ROM checksum self-test: passed (0x24c9f043).
Receiver lock-up workaround activated.

B(eth0) the value of this interface card as per the dmesg of the box is as below

eth0: Intel Corporation 82557 [Ethernet Pro 100], 00:90:27:24:EC:8C, IRQ 15.
Receiver lock-up bug exists -- enabling work-around.
Board assembly 689661-004, Physical connectors present: RJ45 Primary interface chip i82555 PHY #1.
General self-test: passed.
Serial sub-system self-test: passed.
Internal registers self-test: passed.
ROM checksum self-test: passed (0x24c9f043).
Receiver lock-up workaround activated.

B(eth1) this is a wireless aironet 4800 isa card mounted on a RH 7.3 box who's bitrate is set to 11mbps and its status is as below

Every 2s: /bin/cat /proc/driver/aironet/eth1/Status Fri Jan 3 16:20:26 2003

Mode: 1f
Signal Strength: 57
Signal Quality: 125
SSID: solgold
AP:
Freq: 0
BitRate: 11mbs
Driver Version: airo.c 1.5.2 2001/01/22 (Benjamin Reed ) + CISCO EXTENTIONS
Device: PC4800A
Manufacturer: Aironet Wireless Communications
Firmware Version: 3.82
Radio type: 2
Country: 1
Hardware Version: 52
Software Version: 382
Software Subversion: 0
Boot block version: 138

C(eth1) this is a wireless aironet 4800 isa card mounted on a MikroTik 2.4 ver router box who's bitrate is set to 11mbps and its status is as below

[GOLDEN] > interface pc monitor pc1

              quality: 48
             strength: 142
         current-rate: 11Mbit/s
    current-frequency: 2452MHz
         synchronized: yes
           associated: no
                 ssid: solgold
         access-point: CA:03:35:02:16:02
    access-point-name:
         error-number: 0

Regards,
Vishal Gandhi

----- Original Message -----
From: <[email protected]>
To: "Vishal Gandhi Kommineni" <[email protected]>
Sent: Thursday, January 02, 2003 3:24 PM
Subject: Re: [LARTC] speed problem

Quoted 19 lines
Quoted 1 line · Vishal Gandhi Kommineni

Warm wishes for the new year 2003,

Thanks.

Quoted 3 lines

How do i make sure that both the ethernet cards are running in half or full duplex mode, how do i check if weathere an ethernet is in half or full duplex mode

You can use mii-tool. To set eth0 to 100MBit/s halfduplex, run:

# mii-tool -F 100BaseTx-HD eth0

Note: This only works if your card supports this. Or should i say if the driver supports it? Anyway, give it a try.

Quoted 2 lines

Thanks in advance
Vishal Gandhi

Quoted 10 lines

You're welcome, cheers,
Alex.

--
"The most likely way for the world to be destroyed, most experts agree, is by accident. That's where we come in. We're computer professionals. We cause accidents."
-- Nathaniel Borenstein

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

#6 lore raw
In reply to Vishal Gandhi Kommineni

I dont see any collisions when make ifconfig, i feel the problem is in forwarding the packets by the linux box inbetween the interfaces, it is here that there is lot of packet loss,

Do you think it to be so

Thanks & Regards,
Vishal Gandhi

----- Original Message -----
From: "Lars Gaarden" <[email protected]>
To: "LARTC" <[email protected]>
Sent: Thursday, January 02, 2003 4:52 PM
Subject: Re: [LARTC] speed problem

Quoted 3 lines
Quoted 2 lines · Vishal Gandhi Kommineni

Warm wishes for the new year 2003,
thanks for the reply, i found that there are collission when i use a hub

in

Quoted 2 lines
Quoted 2 lines

between,
How do i make sure that both the ethernet cards are running in half or

full

Quoted 21 lines
Quoted 2 lines

duplex mode,
how do i check if weathere an ethernet is in half or full duplex mod

Could you provide the output from 'ifconfig' and 'route -n' on A, B and C? This is either a reverse DNS, duplex mismatch or routing problem.

To see if your are full or half duplex, see if ifconfig reports any collisions on the eth interface - on full duplex you never get any collisions.

If it takes a long time from you try to connect until it starts downloading, you have a DNS problem.

If the download is incredibly slow and one or more of the network interfaces on A, B or C report a huge amount of collisions, you have a duplex problem.

Routing problems usually result in not being able to connect at all, which doesn't seem to be the case here.

Quoted 9 lines

--
The content of this message can either be used according to normal copyright law (including quoting and personal copies) or by the terms of the GNU Free Documentation License http://www.gnu.org/licenses/fdl.html

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

#7 lore raw
In reply to Vishal Gandhi Kommineni

There are collisions in the eth0 interface while there are no collisions in the eth1 wireless interface

I have tried the mii-tool as suggested by [email protected] it did help to some extent but I was not successful in the overall situation I would like to elaborate the setup in more detail

A(eth0) <----->distance 2mts <---> (intel Switch 550T) <---->distance30mts (hub) distance 20mts <--->B(eth0) (eth1 wireless card wan) < ~ ~ ~ ~ ~ > (eth1 wireless card wan) (eth0) <---->distance 2mts (hub) <------>distance 10mts (PC)

A(eth0) the value of this interface card as per the dmesg of the box is as below

eth0: Intel Corporation 82557 [Ethernet Pro 100], 00:90:27:B2:04:82, IRQ 10.
Receiver lock-up bug exists -- enabling work-around.
Board assembly 677173-001, Physical connectors present: RJ45 Primary interface chip i82555 PHY #1.
General self-test: passed.
Serial sub-system self-test: passed.
Internal registers self-test: passed.
ROM checksum self-test: passed (0x24c9f043).
Receiver lock-up workaround activated.

B(eth0) the value of this interface card as per the dmesg of the box is as below

eth0: Intel Corporation 82557 [Ethernet Pro 100], 00:90:27:24:EC:8C, IRQ 15.
Receiver lock-up bug exists -- enabling work-around.
Board assembly 689661-004, Physical connectors present: RJ45 Primary interface chip i82555 PHY #1.
General self-test: passed.
Serial sub-system self-test: passed.
Internal registers self-test: passed.
ROM checksum self-test: passed (0x24c9f043).
Receiver lock-up workaround activated.

B(eth1) this is a wireless aironet 4800 isa card mounted on a RH 7.3 box who's bitrate is set to 11mbps and its status is as below

Every 2s: /bin/cat /proc/driver/aironet/eth1/Status Fri Jan 3 16:20:26 2003

Mode: 1f
Signal Strength: 57
Signal Quality: 125
SSID: solgold
AP:
Freq: 0
BitRate: 11mbs
Driver Version: airo.c 1.5.2 2001/01/22 (Benjamin Reed ) + CISCO EXTENTIONS
Device: PC4800A
Manufacturer: Aironet Wireless Communications
Firmware Version: 3.82
Radio type: 2
Country: 1
Hardware Version: 52
Software Version: 382
Software Subversion: 0
Boot block version: 138

C(eth1) this is a wireless aironet 4800 isa card mounted on a MikroTik 2.4 ver router box who's bitrate is set to 11mbps and its status is as below

[GOLDEN] > interface pc monitor pc1

              quality: 48
             strength: 142
         current-rate: 11Mbit/s
    current-frequency: 2452MHz
         synchronized: yes
           associated: no
                 ssid: solgold
         access-point: CA:03:35:02:16:02
    access-point-name:
         error-number: 0

Regards,
Vishal Gandhi

----- Original Message -----
From: "Martin A. Brown" <[email protected]>
To: "Vishal Gandhi Kommineni" <[email protected]>
Cc: "LARTC" <[email protected]>
Sent: Thursday, January 02, 2003 9:49 PM
Subject: Re: [LARTC] speed problem

Quoted 3 lines

Vishal,

: > the first thing i'd like to check is if the ethernet cards are

running

Quoted 1 line

: > in full- or half-duplex mode. AFAIK running one card in half-duplex

mode

Quoted 83 lines

: > and the other trying to force full-duplex mode can lead to lots of : > collisions, so this could be a reason for slow speed.

I agree with Ulrich, and apparently that may have been your problem.

: thanks for the reply, i found that there are collission when i use a : hub in between, How do i make sure that both the ethernet cards are : running in half or full duplex mode, how do i check if weathere an : ethernet is in half or full duplex mode

You may find that one of the following tools helps you to determine what your ethernet card has negotiated with its peer:

- ethtool ( the tool which will probably continue to be maintained and supported under linux )
http://www.gnu.org/directory/sysadmin/Monitor/ethtool.html

- mii-diag, aka mii-tool, ( one of these two is usually available on common distributions available today )
http://freshmeat.net/projects/mii-diag/

-Martin

Here's some example output from mii-diag and mii-tool:

# mii-tool --help
usage: mii-tool [-VvRrwl] [-A media,... | -F media] [interface ...]
       -V, --version               display version information
       -v, --verbose               more verbose output
       -R, --reset                 reset MII to poweron state
       -r, --restart               restart autonegotiation
       -w, --watch                 monitor for link status changes
       -l, --log                   with -w, write events to syslog
       -A, --advertise=media,...   advertise only specified media
       -F, --force=media           force specified media technology
media: 100baseT4, 100baseTx-FD, 100baseTx-HD, 10baseT-FD, 10baseT-HD,
       (to advertise both HD and FD) 100baseTx, 10baseT
# mii-tool eth0
eth0: negotiated 100baseTx-FD flow-control, link ok
# mii-diag --help
Usage: mii-diag [-aDfrRvVw] [-AF <speed+duplex>] [--watch] <interface>.

Show the link status of the specified network interface. The default interface is "eth0".

The common usage is
mii-diag eth0
Frequently used options are
-A --advertise <speed|setting>
-F --fixed-speed <speed>

        Speed is one of: 100baseT4, 100baseTx, 100baseTx-FD, 100baseTx-HD,
                         10baseT, 10baseT-FD, 10baseT-HD

-s --status Return exit status 2 if there is no link beat.

Less frequently used options are
-a --all-interfaces Show the status all interfaces
(Not recommended with options that change settings.)

   -D  --debug
   -g  --read-parameters        Get driver-specific parameters.
   -G  --set-parameters PARMS   Set driver-specific parameters.
       Parameters are comma separated, missing elements retain existing
value.
   -M  --msg-level LEVEL        Set the driver message bit map.
   -p  --phy ADDR               Set the PHY (MII address) to report.
   -r  --restart        Restart the link autonegotiation.
   -R  --reset          Reset the transceiver.
   -v  --verbose        Report each action taken.
   -V  --version        Emit version information.
   -w  --watch          Continuously monitor the transceiver and report

changes.

This command returns success (zero) if the interface information can be read. If the --status option is passed, a zero return means that the interface has link beat.

# mii-diag eth0
Basic registers of MII PHY #1:  3000 7809 02a8 0154 05e1 0000 0000 0000.
 Basic mode control register 0x3000: Auto-negotiation enabled.
 Basic mode status register 0x7809 ... 7809.
   Link status: not established.
   End of basic transceiver information.
Quoted 4 lines

--
Martin A. Brown --- SecurePipe, Inc. --- [email protected]

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

#8 lore raw
In reply to Vishal Gandhi Kommineni

try to keep the frequency of the 2 wireless Aironet cards as far appart as possible. then run the test again.

set one card for CH 1 ( or 2412 Mhz ) and the other to CH 11 ( or
2460 Mhz )

-Dan Kramarsky

Speedband, Inc
Los Angeles, CA.

Wireless Internet Service

www.speedband.com

==Dan Kramarsky
RF Engineer
SpeedBand, Inc.
=- Wireless Internet Service Provider -
phone: 909-551-7358
web: www.speedband.com

__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

#9 lore raw
In reply to Vishal Gandhi Kommineni

Thank you sir for your suggestion, is there any tool throught which we can impose the selected frequency on the aironet card.

Because even if i select channel 1 in the Config file, it is setting to channel 9 any directions in this angle are welcomed.

REgards,
Vishal

----- Original Message -----
From: "Dan Kramarsky" <[email protected]>
To: <[email protected]>
Sent: Friday, January 03, 2003 11:28 PM
Subject: Re: [LARTC] speed problem

Quoted 22 lines

try to keep the frequency of the 2 wireless Aironet cards as far appart as possible. then run the test again.

set one card for CH 1 ( or 2412 Mhz ) and the other to CH 11 ( or
2460 Mhz )

-Dan Kramarsky

Speedband, Inc
Los Angeles, CA.

Wireless Internet Service

www.speedband.com

==> Dan Kramarsky
RF Engineer
SpeedBand, Inc.
=- Wireless Internet Service Provider ->
phone: 909-551-7358
web: www.speedband.com

Quoted 8 lines

__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/