Opened 11 years ago

Last modified 11 years ago

#3493 closed Bug report

Downloads a much bigger file than exists

Reported by: corylink Owned by:
Priority: normal Component: FileZilla Client
Keywords: Cc: corylink, Alexander Schuch, Tim Kosse
Component version: Operating system type:
Operating system version:

Description

Version 3.0.9.2
Windows XP SP2

Connected to a buddies FTP server to download a couple of .rar files. One file was 85MB and the other was 76MB. Everything looked fine until the file reached the expected size. At that point, the time remaining stayed at 0 and the file continued to download. It kept downloading and around the point where it was approximately 150MB, I copied the file to another folder to attempt to open it. I opened it, and the contents looked fine. It continued to download until it was 156MB and then produced the following error: The data is invalid.
Error: Critical error

Same exact thing with the second file (except that one went until 145MB).

I extracted the file one more time after the download completed and again, the contents seemed fine, just the file itself was about twice as big as it was supposed to be.

Attachments (1)

log.txt (8.8 KB) - added by corylink 11 years ago.
Copy of log from log window [PRIVATE]

Download all attachments as: .zip

Change History (3)

Changed 11 years ago by corylink

Attachment: log.txt added

Copy of log from log window [PRIVATE]

comment:1 Changed 11 years ago by Alexander Schuch

Was the file on the server fully stored already? By this I mean that if the server admin was copying the file from another media to the server's shared FTP directory, and a client then connects and fetches the directory list while the file is still being copied to the shared directory, the FTP client gets the wrong file size - current file size of the copy process, which is less than the actual size of the complete file.

Your connection to the server was quite bad, so the connection dropped from time to time and FileZilla resumed the transfer after reconnect. In order to resume a file, FileZilla sends the RETR command followed by a file offset from which on it wants to read/download data. The offset is based on how many byte of the file were downloaded already. My guess for the critical error now is that the file offset requested by FileZilla was behind the end of file, so the server couldn't send any data and bailed out with an error. So something went very wrong, but I have no idea what exactly.

comment:2 Changed 11 years ago by Tim Kosse

This is a well known bug in Microsoft's FTP server. It claims to support resume, accepts the REST command yet it has no effect.

Please switch to a proper server.

Note: See TracTickets for help on using tickets.