On 7/31/19 3:21 PM, Julie Nishimura wrote:
> Hello postgres folks,
>
> We're tossing around the idea of upgrading a replicated postgres cluster
> (37 dbs) by breaking the replication, adding different size (larger)
> data disks to the hot-spare, then turning replication back on, letting
> it fully populate, then breaking replication, making the standby the
> primary, upgrade the disks on the other system, bring it back up,
> replicate backwards until fully replicated then failing-back to the
> original primary. Is this feasible?
To make a determination more information would be helpful:
1) What is the method of replication?
2) What is the expected downtime for the disk upgrades?
3) Dependent on 1). Is there sufficient storage to hold the data until
it can be replicated back?
4) How is the data on the existing disks going to get to the new disks?
>
> Our current size is 22 tb, and it is 97% full
> (PostgreSQL 9.6.2 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu
> 5.3.1-14ubuntu2) 5.3.1 20160413, 64-bit)
>
>
> Thank you for your suggestions
--
Adrian Klaver
adrian.klaver@aklaver.com