[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Checksum mismatch error becoming critical

From: Ben Collins-Sussman <sussman_at_collab.net>
Date: 2005-05-26 01:59:03 CEST

On May 25, 2005, at 4:40 PM, Joshua Jensen wrote:

>>> In my experience, such a checksum mismatch symptom
>> indicates a network
>>> problem, not a Subversion problem.
>> Exactly, this is the point of svn's checksums. They're meant
>> to detect errors in the things *outside* of Subversion.
>> Almost every checksum error report we've ever seen has been
>> the result of bad disk, bad ram, or flaky network.
> I see this often, too. Why aren't the network packets
> checksum'ed? If they
> are, how is valid data received from the network being applied to
> the .svn
> working copy such that the checksum fails?

The packets themselves aren't checksummed; TCP/IP already does that
to some extent, no?.

But whenever subversion sends file data over the network (a binary
diff) in either direction, a checksum is sent for the final file to
be constructed with the diff. If the checksum doesn't match the
reconstructed file, then you see an error. The change is *not*
applied to the working copy or repository.

To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu May 26 02:04:07 2005

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.