What is the nature of the corruption? Are the files correct up to a certain point, and then simply end, suggesting that the data stream got cut off, or are there missing bytes within the file (so the resultant length is shortened), or are some bytes changed (so the final file is the right length but some of the contents are incorrect)?
The length is always correct.
ZIP files: CRC Error
EXE files: errors when running them (if standalone programs) or when installing them (like Adobe Acrobat Reader 6).
What happens if you download the same file across the same leased line but bypassing the Squid proxy?
I did, and as I mentioned before, nothing changed. My parent proxy (which was my main proxy, until I dediced to put another local proxy here at my offce) seems still to cause corrupted downloads...
What type of file/s get corrupted? Plain text? HTML? Binaries (images, executables, arbitrary data)?
Mainly BIG files ( about every file bigger than 4/5MB), therefore ZIPs and EXEs.
What is the MIME content type of the downloads which get corrupted? Text/plain? Text/html? Application/octet-stream?
The last one, "application/octet-stream".
What do the Squid access.log files on both your proxy and the parent proxy tell you about the files which arrive corrupted?
I can't get the parent proxy's log files, but mine is like:
1066382936.928 94247 10.167.211.11 TCP_MISS/200 18033719 GET http://ardownload.adobe.com/pub/adobe/reader/win/6.x/6.0/ita/AdbeRdr60_ita_full.exe boniforti FIRST_PARENT_MISS/proxy.reteunitaria.piemonte.it application/octet-stream
Hope some of these questions point in a helpful direction.....
Regards,
Antony.
Thank you...
