Chronological Thread 

< Chronological >      < Thread >

---

Hi,

The new teaching instance has been prepared; a final data migration using
data from existing teaching instance will be done in about twelve hours with
the rollover from the existing instance to that new instance done immediately
after. There will be a brief service interruption while that is done.

The staging instance has been rebuilt, with its data restored from production
backups for the current staging deployment.

Both these instances are now on separate VMs with more RAM allocated, and are
running on the same production profile as the production server, so hopefully
going forward these will be as stable as the production instance. Naturally,
this staging (and teaching, to an extent) will be used for gauging the actual
stability of this completely new server setup that is produced with a new set
of mostly automated deployment scripts. If this new setup doesn't crash over
the next month or two I will then action the same deployment for the
production instance.

Cheers,
Tommy.

On 20/03/18 17:20, Tommy Yu wrote:
> Hi,
>
> Due to various delays, the staging instance was not refreshed with restored
> data last month. Going forward, please expect that the data on the staging
> instance be wiped periodically without warning. On the other hand, wiping
> of the data on the teaching instance will have warnings in place.
>
> The staging instance will be fully wiped and restored using the VM
> installation scripts that are now done for this purpose in the next few
> hours.
>
> Cheers,
> Tommy.
>
> On 14/02/18 19:50, Tommy Yu wrote:
>> Hi,
>>
>> The PMR staging instance <https://staging.physiomeproject.org> will be
>> refreshed by end of this month (26 February), as there are some drastic
>> backend changes that need to be effectuated to ease deployment and
>> maintainability of the overall infrastructure. The teaching instance
>> <https://teaching.physiomeproject.org> will be rebuilt by end of next
>> month pending the stability of the new staged installation, with the
>> production server be rebuilt shortly after that.
>>
>> Currently, the teaching PMR instance is simply an alias to the staging
>> instance. This was not an issue when both teaching and staging instances
>> received little to no usage, but now that the teaching instance is being
>> used, the debug settings for the staging instances results in performance
>> issues that would not be present otherwise.
>>
>> The current plan follows:
>>
>> - rebuild the staging server, user data could be in any state (depending
>> on the nature of the tests/deployment exercise to be executed, it can be
>> either a full restore from backup or a fresh installation with minimum
>> user data).
>> - rebuild the teaching server, user data to be copied from existing
>> teaching server (this does not exclude future data wipe/partial restore
>> from production backups, as user data on teaching is never backed up).
>>
>> Once the teaching server instance migration is successful, the same method
>> will be used to migrate the production instance. Do note that the plans
>> for teaching instance may change, so like before please treat that as a
>> transient data store rather than some permanent fixture.
>>
>> Regards,
>> Tommy.
>>
>

---

• Re: [cellml-discussion] PMR staging/teaching server data wipe/restructuring, Tommy Yu, 03/21/2018
  • Re: [cellml-discussion] PMR staging/teaching server data wipe/restructuring, Tommy Yu, 03/23/2018

---

Archive powered by MHonArc 2.6.18.