On 11/06/14 17:43, Kurt Roeckx wrote:
On Wed, Jun 11, 2014 at 11:13:03AM +0100, Ian Chard wrote:
I downgraded openssl and libssl1.0.0 to 1.0.1e-2+deb7u9 this morning,
and my job has been running for two hours without incident. Previously
it was failing after 15 minutes or so.
I've got someone else reporting the same issue, also with
postgres. In that case it was from a box running wheezy
(with 1.0.1) connecting to a server running squeeze (with 0.9.8).
You also mentioned that it's an upgrade, and maybe postgres is
still using libssl0.9.8?
Please note that it's not because you have libssl1.0.0 installed
that you're also really using it. postgres needs to be linked
with that version. I suspect it's still using the 0.9.8 version.
I had a look at which libraries were in use by the job that was having a
problem, and for some reason it has both 0.9.8 and 1.0.1 loaded.
I've also had another report from a different host (accessing the same
DB server); that one is using the psycopg2 library which only uses 1.0.1:
ian@milo:~$ ldd /usr/lib/pyshared/python2.7/psycopg2/_psycopg.so|grep -i ssl
libssl.so.1.0.0 = /usr/lib/x86_64-linux-gnu/libssl.so.1.0.0
(0x7ff78967a000)
I've downgraded openssl on that machine too in the hope that it'll fix
the problem in the same way (the symptoms, and the CCS message, were the
same).
Cheers
- Ian
--
Ian Chard i...@mysociety.org
mySociety systems administrator http://www.mysociety.org/
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
