Tom Hughes

http://dba.stackexchange.com/questions/59530/streaming-replication-between-postgresql-9-2-and-9-3 appears to confirm.

The question is, what are our chances of doing an in-place upgrade rather than a dump+reload, and if we can do that how long might it take...

I've had issues in the past with 3GL functions, though those can usually be worked around I think Fedora at least also makes a copy of the database when using...

Yes I know (it's not the default on Ubuntu, in fact dump and reload is the default...) and we've already tested it on the (much larger) main database.

Do we have the right metrics for that? Trying to under the histogram and *ile metrics always makes my head hurt...

As far as I know the only places we have dash boards for response time is for the main web site and nominatim. The main web site is just based...

As an example this should do p95 for nominatim by request type: ``` histogram_quantile(0.95, sum by (type, le) (rate(nominatim_request_duration_seconds_bucket[$__rate_interval]))) ```

It's not hard, it just needs notice so we can schedule it, ideally for a Sunday morning or failing that for a Saturday morning.

Oh I can reproduce it, though only when using the chinese language interface as you say - switch to english and it is ok. No idea what we can do...

Given how traumatic the last upgrade was I'm in no hurry to do another without some sort of guarantees that it isn't going to break everything.