Direct uploads are slower with v3.1.x

JCDinPGH · 03-29-2005, 08:05 PM

I've been trying several of the non-public betas and also now public beta 3 of v3.1, and they all seem to be much slower when doing direct uploads. I'm not sure if this is related to this thread but it seems similar, although I'm not on a gigabit network. (http://forum.flashfxp.com/showthread...&threadid=5597).
I'm on a standard 10/100 network but the differences in speeds when uploading between v3.0.2 build 1045 and v3.1.12 build 1075 is pretty extreme like 150k-200k/sec difference when sending to a site with max bandwidth of 450k/sec. With v3.0.2, I can max out the connection, with v3.1.12, I can only get about half the max bandwidth. I've tried tweaking the settings and it didn't seem to change anything. I have upload packet size on v3.0.2 set to 4096 which maxes out the connection and tried setting v3.1.12 to send and receive of both 4k, but it didn't change anything.
Any suggestions?
John

bigstar · 04-03-2005, 05:11 AM

Are you using passive mode or port mode?

And are you using the same mode in v3.0 and v3.1?

JCDinPGH · 04-04-2005, 09:19 AM

Quote:

Originally posted by bigstar
Are you using passive mode or port mode?

And are you using the same mode in v3.0 and v3.1?

I am using PASV for both versions. When I installed v3.1 beta, I copied all the files from my normal install dir into a separate directory, and then installed v3.1 beta over top of the new directory so it retained my previous settings. PASV is currently on.
John

bigstar · 04-04-2005, 11:51 AM

As a test rename the "flashfxp.ini" file in your 3.1 install to something else say "flashfxp.ini.bak", this will reset all of the settings to default. Please let me know if the default settings provide better results.

If the above doesn't work try this.. If your internet configuration allows it, try turning off passive mode in FlashFXP.

JCDinPGH · 04-04-2005, 03:17 PM

Quote:

Originally posted by bigstar
As a test rename the "flashfxp.ini" file in your 3.1 install to something else say "flashfxp.ini.bak", this will reset all of the settings to default. Please let me know if the default settings provide better results.

If the above doesn't work try this.. If your internet configuration allows it, try turning off passive mode in FlashFXP.

I renamed the flashfxp.ini and the speeds were the same. I am forced to use PASV because I am behind a firewall that I don't control so I can't control the port options. The ftp daemon I'm sending to is IOFTPD v5-8-5r although I'm not sure that matters. I am also going to try sending something from one server to another internally in our network where I *can* turn off PASV and see what happens.
Stay tuned,
J.

JCDinPGH · 04-04-2005, 08:28 PM

Quote:

Originally posted by swspjcd
I renamed the flashfxp.ini and the speeds were the same. I am forced to use PASV because I am behind a firewall that I don't control so I can't control the port options. The ftp daemon I'm sending to is IOFTPD v5-8-5r although I'm not sure that matters. I am also going to try sending something from one server to another internally in our network where I *can* turn off PASV and see what happens.
Stay tuned,
J.

I did some transfers on our local 100Mb network and the differences between v3.02 and 3.1beta were about 1300k/sec with 3.1beta being the slower. PASV was off for both tests. Both had packet sizes of 8k. I can get more detailed tests if you like.
J.

bigstar · 04-04-2005, 09:44 PM

I believe the latest 3.1 beta has this option.. when you select the buffer size scroll all the way to the top and select "default" as the value, this is the value used by 3.0, See if this makes a difference.

JCDinPGH · 04-05-2005, 09:04 AM

Quote:

Originally posted by swspjcd
I did some transfers on our local 100Mb network and the differences between v3.02 and 3.1beta were about 1300k/sec with 3.1beta being the slower. PASV was off for both tests. Both had packet sizes of 8k. I can get more detailed tests if you like.
J.

I changed the packet size to DEFAULT on v3.1beta and got about 7600k/sec. When I changed the packet size to 8K on v3.02, I got roughly the same speeds. If I change the packet size on v3.02 to 4096, I get about 1200k more per second although changing the packet size on v3.1beta to 4096 only gets about 5100k/sec. Even weirder. The fluctuations between my speed tests last night and today are expected since there is more network traffic during the day. The speed differences seem relative though.
J.