Re: [GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-07 Thread Peter Geoghegan
On Thu, Aug 7, 2014 at 9:46 AM, Bruce Momjian wrote: > We could walk the index looking for inconsistent btree splits, e.g. the > split doesn't match the ordering returned by the existing collation > functions. I'm not sure I follow. I don't think that a tool like my btreecheck tool will necessari

Re: [GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-07 Thread Bruce Momjian
On Thu, Aug 7, 2014 at 03:07:04PM +, Matthew Kelly wrote: > We are currently running with the en_US.UTF-8 collation. It was a decision > made long ago, and seeing as we never actually rely on the sort order of > internationalized strings (other than for stability, apparently), we have > ne

Re: [GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-07 Thread Matthew Kelly
We are currently running with the en_US.UTF-8 collation. It was a decision made long ago, and seeing as we never actually rely on the sort order of internationalized strings (other than for stability, apparently), we have never had any motivation to change this practice. Some way of versioning

Re: [GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-06 Thread Peter Geoghegan
On Wed, Aug 6, 2014 at 6:30 PM, Tatsuo Ishii wrote: > Another idea could be having our own collation data to isolate any > changes from outside world. I vaguley recall this had been discussed > before. That's probably the best solution. It would not be the first time that we decided to stop relyi

Re: [GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-06 Thread Tatsuo Ishii
> Over time, collation order will vary: there may be fixes needed as > more information becomes available about languages; there may be new > government or industry standards for the language that require > changes; and finally, new characters added to the Unicode Standard > will interleave with th

Re: [GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-06 Thread Peter Geoghegan
On Wed, Aug 6, 2014 at 5:11 PM, Bruce Momjian wrote: > No surprise; I have been expecting to hear about such breakage, and am > surprised we hear about it so rarely. We really have no way of testing > for breakage either. :-( I guess that Trip Advisor were using some particular collation that

Re: [GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-06 Thread Bruce Momjian
On Wed, Aug 6, 2014 at 09:24:17PM +, Matthew Kelly wrote: > The following is a real critical problem that we ran into here at TripAdvisor, > but have yet figured out a clear way to mitigate. > > TL;DR: > Streaming replicas—and by extension, base backups—can become dangerously > broken > when

[GENERAL] The dangers of streaming across versions of glibc: A cautionary tale

2014-08-06 Thread Matthew Kelly
The following is a real critical problem that we ran into here at TripAdvisor, but have yet figured out a clear way to mitigate. TL;DR: Streaming replicas—and by extension, base backups—can become dangerously broken when the source and target machines run slightly different versions of glibc.