On 2/27/23 08:49, Dávid Suchan wrote:
Reply to list
Ccing list for real this time.
1) I used \l+ in psql and then counted rows- millions were missing
\l lists databases.
Are you saying there are millions of database?
Otherwise what rows where you counting?
3) nothing at all, everything was "success"
4) I did not, I presume it is there, the question is why only 700 mb was
transferred
If you have not connected how could you do the \l and row count?
5) would it be inside main pg log? Or some special one?
Dňa po 27. 2. 2023, 17:14 Adrian Klaver <adrian.klaver@xxxxxxxxxxx
<mailto:adrian.klaver@xxxxxxxxxxx>> napísal(a):
On 2/27/23 07:44, Dávid Suchan wrote:
> Hello, I tried upgrading pg db from version 9.6 to 14 by using
> pg_upgradecluster command. I freshly installed pg 14 -> ran
> pg_dropcluster 14 main --stop -> and then upgraded using
> pg_upgradecluster 9.6 main.
> After a successful prompt finished, I checked the database and
the size
> went from originally 20gb (in 9.6) to 700~ mb (in 14) while the disk
> space available shrank by about 2gb meaning that there is still
the 20gb
> of data. I tried the entire process twice (since I had created an
AWS
1) How did you measure the size of the database clusters?
2) pg_upgrade will not remove the old cluster automatically so it not
surprising the overall disk usage increased.
3) Did you see any messages at end of upgrade mentioning issues?
4) Have you connected to new cluster to see if everything is there?
5) Does the Postgres log provide any relevant information?
> EC2 snapshot for this) and the result was the same.
> Is my solution to migrating old pg version to the new one wrong?
Before
> this I tried the same process with around 300mb of data and all
of that
> transferred successfully. If I did not understand the
pg_upgradecluster
> command, what would be the best practice when upgrading pg
version with
> huge amounts of data(could be a terabyte)?
--
Adrian Klaver
adrian.klaver@xxxxxxxxxxx <mailto:adrian.klaver@xxxxxxxxxxx>
--
Adrian Klaver
adrian.klaver@xxxxxxxxxxx