Tue May 4 06:04:11 PDT 2010
- Previous message: [Slony1-bugs] UTF-8 encoding error...problems with 2.0.3
- Next message: [Slony1-bugs] An old event not confirmed: A possible bug?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Andrew Eross wrote: > Hi folks, > > My company (and I) run Slony for 4 of our production databases, and have > been using it for years. > > All 3 of our database servers are consistently running Postgres 8.3.10 > on Ubuntu, all databases created with ENCODING = 'UTF8'. > > I recently attempted to upgrade from 2.0.2 to 2.0.3, which went > (seemingly) quite smoothly, the upgrade caused itself went fine, > following the standard instructions. Your problems are caused by the upgrade. 2.0.3 has a bug. See this thread http://lists.slony.info/pipermail/slony1-general/2010-April/010596.html A fix has already been committed to cvs > > However, afterwards our replication completely broke, and I started > seeing lots of UTF8 encoding errors reported in the slony log files for > 2 of our databases that had been previously replicating just fine. > > Lots of these messages: "ERROR: invalid byte sequence > for *encoding* "UTF8":" > > I've seen some posts out there in the world with the same problem, > probably from a different cause. e.g., this > guy: http://old.nabble.com/encoding-problems-td28383260.html > > Same thing happened to me, and it was not a Postgres versioning issue, > this started occurring only after I upgraded to 2.0.3. > > I eventually wrote a tool to manually traverse the entire database in > question and select/update every row after running it through iconv to > be sure there was no encoding issues. I'll open source the tool later, > it was a handy thing to make anyways. > > After thatI tried dropping replication and starting from scratch with > the refreshed data, still resulted in the same error. > > Anyways, how I've ended up fixing it is I've concluded the problem only > occurs with Slony 2.0.3 ... I have now re-installed 2.0.2 and re-created > all the replication from scratch...and poof no more UTF8 encoding errors. > > Sorry, I wish I had the log files, but they got wiped out. I just wanted > to ask if this was a known issue? If not, I will endeavour to go > re-create the bug again and get the logs this time. > > PS > Just a side note -- anecdotally -- separately I saw some other errors in > the log files not related to UTF8 encoding, just a general INSERT error > because it looked like Slony was trying to run an SQL query either with > a missing double-quote or with-out properly escaping the string first > --- but, sorry, again, someone deleted our logs, so I'll try to dig that > out again too. > > Cheers! > > Andrew Eross > CTO > Locatrix Communications > +61.405.761.614 > eross at locatrix.com <mailto:eross at locatrix.com> > > > ------------------------------------------------------------------------ > > _______________________________________________ > Slony1-bugs mailing list > Slony1-bugs at lists.slony.info > http://lists.slony.info/mailman/listinfo/slony1-bugs -- Steve Singer Afilias Canada Data Services Developer 416-673-1142
- Previous message: [Slony1-bugs] UTF-8 encoding error...problems with 2.0.3
- Next message: [Slony1-bugs] An old event not confirmed: A possible bug?
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-bugs mailing list