Re: Postgres 12.1 : UPPER() in WHERE clause restarts server

Adrian,

Old production server was postgres 9.6 with Mac0S 10.9 so much older
than the “new” server. ( Now MacOS 10.14 Postgres 12.1 )
After sudden restart of the cpu we started having issues, part of the
data that is lost TOAST. . . and we also started having issues when
TRUNCATING certain tables. The tabel where Nick reported the
“UPPER”-issue was not “involved”

The data we restored today on the “new” server was from before the
crash, and in our opinion “healthy”, didn’t run into any issue
importing it.

Never had issues before, RAID system wasn’t giving any warnings
neither. We still presume an hardware failure, but haven’t been able
to figure out what exactly.

I’ll provide a log tomorrow in the morning CET.

Regarding the stack trace we’ll dig in to it to see what we can come
up with

¿ is downgrading possible ? We need to have a working system by
tomorrow evening CET.

Thanks in advance,

Marc

On 8 Feb 2020, at 22:18, Adrian Klaver wrote:

> On 2/8/20 12:28 PM, Marc wrote:
>> Adrian,
>>
>> Everything was a clean install ( MacOS Mojave and Postgres )
>>
>> Export and import were done with the latest version of PGAdmin.
>>
>> Please advise if we can provide you with anything ( logging etc . . .
>> )
>>
>>
>> Is there a possibility to downgrade to version 11 ?
>
> At this point hard to tell whether this a version issue or something
> else. Probably best not to introduce too many more moving parts at
> this time.
>
> Questions:
>
> 1) The OS and version you mention above is the same as on the
> production server below?
>
> 2) What where the exact issues you had on the production server?
> The actual error messages would be very helpful.
>
> 3) Getting a stack trace of the UPPER(), LOWER() issue would be nice.
> For more info on that:
>
> Not that familiar with MacOS, so not sure if the Linux/BSD steps would
> apply or not, still:
>
> https://wiki.postgresql.org/wiki/Generating_a_stack_trace_of_a_PostgreSQL_backend
>
>
>
>
>
>>
>> We upgraded over the weekend because we experienced a crash on our
>> production server with “toast” issues as result.
>>
>> Thanks in advance,
>>
>> Marc
>>
>>
>
>
> --
> Adrian Klaver
> adrian(dot)klaver(at)aklaver(dot)com

ARC - your Apple service partner