Re: Possible corrupt index?

On 4/17/19 7:28 AM, Zahir Lalani wrote:
>
>> Any changes occur between the time it worked and the time it did not?
>> Say a crash, change in schema, new data added and so on.
>
> Not as far as we can tell - one of my DB team noticed it day before yesterday - there were no tell tale signs of any issues but we have seen issues yesterday with certain records not being found

Does that including looking at the logs?

>
>> Per below the version 10 servers use en_GB.UTF_8 and the production one en_US.UTF-8.
>> Why the difference?
>> And did the production used to be en_GB.UTF_8?
>
> All servers used to be en_US as I think that was the default and had been setup early in our use of PG. I suspect our devops guy corrected this during the upgrade.

Did that extend to fiddling with the locale on the production machine?

>
>> Have you logged into the production instance using psql and tried the queries to see if they work?
> Not via psql, but as I said we saw issues from our app layer. I will do a similar test with psql and let you know
>
>
> Z
>

--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com