On Wed, 15 Apr 1998, Oskar Pearson wrote:
...
> Squid-1.2 now includes data like this in the actual cache objects...
> including the URL... it throws it away if the on-disk object URL and
> the one that the client is actually trying to get is different...
That's a step in the right direction, but will not catch 100% of common
data corruption. It's entirely possible that the header of the file is
correctly updated, but not all of the contents. Unless squid makes a
checksum/CRC over the entire file (which would be a big performance hit),
it might not notice that the last few blocks of the file are garbled.
This is especially likely for the case originally cited, which is that
several of the files on the cache partition were coming back as having
duplicate blocks.
-ben
ps: Sorry for ranting on about this, but I believe that ISPs should use
web caches *and* ensure that the data is as consistent as possible.
There's nothing more frustrating than being transparently proxied and not
getting valid data back from an entirely avoidable cause.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu