Skip to main content

[archive] C$COPY and zlib

  • May 4, 2007
  • 4 replies
  • 0 views

[Migrated content. Thread originally posted on 02 May 2007]

Hello everyone,
I've a TC (7.0.0) on a Win2003 server that heavily uses [@DISPLAY] C$COPY to transfer files from client to server from server to client again, with big files (more than 30 Mb each) sometimes it slows down "dripping" byte per byte from source to destination, going into an unreasonable transfer time.
There's no particoular network traffic to generate a bottleneck slowdown.
On runtime is setted:
AGS_SOCKET_COMPRESS ZLIB
Has someone's experienced some similar situation?
bye, Giovanni.

4 replies

[Migrated content. Thread originally posted on 02 May 2007]

Hello everyone,
I've a TC (7.0.0) on a Win2003 server that heavily uses [@DISPLAY] C$COPY to transfer files from client to server from server to client again, with big files (more than 30 Mb each) sometimes it slows down "dripping" byte per byte from source to destination, going into an unreasonable transfer time.
There's no particoular network traffic to generate a bottleneck slowdown.
On runtime is setted:
AGS_SOCKET_COMPRESS ZLIB
Has someone's experienced some similar situation?
bye, Giovanni.
Hello everyone,
according to our local Acu tech support the behaviour we experienced is due to some socket communication buffer bugs: situation was partially pathched with ECN 3377 on 7.2.0 and was again patched on 8.0. Patch of 8.0, they say, can't be ported on 7.2.x :mad: : maybe here in Italy we've to take out money for major release...?
bye Giovanni

[Migrated content. Thread originally posted on 02 May 2007]

Hello everyone,
I've a TC (7.0.0) on a Win2003 server that heavily uses [@DISPLAY] C$COPY to transfer files from client to server from server to client again, with big files (more than 30 Mb each) sometimes it slows down "dripping" byte per byte from source to destination, going into an unreasonable transfer time.
There's no particoular network traffic to generate a bottleneck slowdown.
On runtime is setted:
AGS_SOCKET_COMPRESS ZLIB
Has someone's experienced some similar situation?
bye, Giovanni.
Some test results: I tried 720 version even if it includes ECN 3377 it didn't resolved our slow transfer speed problems. I think we must wait for 8.0...
Bye giovanni

[Migrated content. Thread originally posted on 02 May 2007]

Hello everyone,
I've a TC (7.0.0) on a Win2003 server that heavily uses [@DISPLAY] C$COPY to transfer files from client to server from server to client again, with big files (more than 30 Mb each) sometimes it slows down "dripping" byte per byte from source to destination, going into an unreasonable transfer time.
There's no particoular network traffic to generate a bottleneck slowdown.
On runtime is setted:
AGS_SOCKET_COMPRESS ZLIB
Has someone's experienced some similar situation?
bye, Giovanni.
I found the same issue with Version 6.1.0. It only was a problem when copying files from the client to the server (VERY SLOWLY). since we already had FTP accounts in use on that server I changed my code to FTP the file(s) instead and solved the problem, A lot more coding but it was 1000 times faster.

[Migrated content. Thread originally posted on 02 May 2007]

Hello everyone,
I've a TC (7.0.0) on a Win2003 server that heavily uses [@DISPLAY] C$COPY to transfer files from client to server from server to client again, with big files (more than 30 Mb each) sometimes it slows down "dripping" byte per byte from source to destination, going into an unreasonable transfer time.
There's no particoular network traffic to generate a bottleneck slowdown.
On runtime is setted:
AGS_SOCKET_COMPRESS ZLIB
Has someone's experienced some similar situation?
bye, Giovanni.
Hi, we're now making the idea to drive out of C$COPY and dive into FTP use: 1) did you consider the use of FTP protocol via command line? (es C$SYSTEM and DOS FTP commands)
2) I look at a message you posted some days ago about problems using FTP and Thin Client (an extra byte added...), did you solved this problems now or not FTP use via Thin must be still considered an unstable technique?
Thanks in Advance, bye Giovanni.