[Pkg-openssl-devel] Bug#751093: Bug#751093: not so sure it's openssl

[Kurt Roeckx]

On Tue, Jun 10, 2014 at 06:11:50PM +0200, Tomasz Majbaum wrote:
> The very similar problem occured on my server machine (which I upgraded
> from Squeeze to wheezy yesterday) with locally executed dump. Backup was
> interrupted every time when it reached one of the biggest tables,
> stopping at approx. 77MB of data. I've gone back to openssl
> 0.9.8o-4squeeze14, but it didn't help.
> 
> Oddly enough, cluster upgrade did the trick (from 8.3 to 9.1) and now it
> works with current openssl.
> 
> My error message was:
> 
> [...]
> pg_dump: restoring data for table "xyz"
> pg_dump: dumping contents of table xyz
> pg_dump: Dumping the contents of table "xyz" failed: PQgetCopyData()
> failed.
> pg_dump: Error message from server: blad SSL: ccs received early
> pg_dump: The command was: COPY a.xyz (x,y,z) TO stdout;
> pg_dump: *** aborted because of error

Can you retry this with libssl1.0.0 1.0.1e-2+deb7u9 instead
of 1.0.1e-2+deb7u10?

The 1.0.1e-2+deb7u10 did some changes related to CCS where
it now gives that error but in older versions didn't always
do it when it should.


Kurt