Hello
We have a fairly large postgresql-9.4 database (currently 70TB and growing
to approx 80TB ) running on Centos 7.
The HW is 48 core Xeon with 180GB of RAM with data on a enterprise grade
SAN storage.
We started feeding it several weeks ago and everything went smoothly until
we hit this issu
d changed after vacuum full, or am I not
understanding something?
Thanks a lot
On Wed, Aug 9, 2017 at 7:57 PM Andres Freund wrote:
> Hi,
>
> On 2017-08-09 10:06:48 +, Peter Hunčár wrote:
> > We started feeding it several weeks ago and everything went smoothly
> until
> >
Hello,
there are currently no transactions whatsoever, the app is paused. I can
even restart the database if needed.
I ran vacuum full, because as I mentioned above it seemed to me that manual
vacuum did not change the relminmxid of a table.
Unfortunately, an upgrade is not an option :(
Hello
it took 10 days to autovacuum the 32TB toast table and after restarting
with previous (but slightly tuned autovacuum) parameters, the database
works just fine.
The data import is almost done, there are just several hundred thousands of
rows to add :)
The next headache will the the backup of
Hi,
we have a table with around 1.6 billion rows having quite lot of big binary
data toasted.
Today we started getting:
WIB > ERROR: invalid page in block 1288868309 of relation base/96031/96201
Which is a toast reltype.
I know that zero_damaged_pages and vacuum (or restore the table from
back
