#6809 Transitioning Ask -- giving Evgeny Fadeev access to the server to do the transition
Closed: Fixed 5 years ago Opened 6 years ago by mattdm.

How do we want to handle this? We've signed a contract and statement of work -- Evgeny, the Askbot developer, is going to take over hosting and day-to-day administration, and work on some specific improvements we've identified.

He says the next thing he needs is to log into the server to transfer data. How do we want to go about this?


If he could make a fas account (I think he might already have one) and upload the key there, we can add him to sysadmin-ask that has perms to login to the servers. We also can make him an admin in the interface as soon as we know what his account is.

Ah, it's efadeev

If you can pass along the above that would be great.

He's using an askbot account for this.

(Can we, by the way, add this account to a "not a human" group?)

ok, so re-reading this he mostly needs the existing data right?

How about we put up a database dump of our existing askbot db and a tar.gz of the existing uploads/ dir. Is there any other things he needs as far as content? We can get those setup and he can do some kind of initial setup/test and if all looks well we could schedule downtime, get the latest db upload and change dns to point to his instance.

Yes, he says that's fine. How do we get that to him?

Could you ask him to join the ticket. It would be helpful to get exactly what he is needing and other items he might need from us.

Metadata Update from @kevin:
- Issue private status set to: False (was: True)

6 years ago

ok. I have added the askbot account to a group so it has access to fedorapeople.org.

We are placing a db dump and uploads tar.gz in ~askbot on fedorapeople.org.

After those are found to be good, we can schedule an outage, shutdown our instance, do a db dump and upload that and then switch dns over.

@kevin those are the files that I need. Found them, thanks.

ok, what are the next steps here?

I guess if those data files look ok, we can schedule a cutover / outage.

We can then take down the service on our end then, do a final dump/collection of upload and get it to @askbot then when ready on the other end, we point dns at whatever IP they give us.

Let us know when would be a good time to schedule this.

Metadata Update from @kevin:
- Issue assigned to kevin
- Issue priority set to: Waiting on External

6 years ago

Hi @kevin , can we schedule this for the coming Thursday? Perhaps morning in your time zone?

Can you give times UTC you are available? Kevin's Thursday is usually booked heavily and do not start until 15:00 UTC. We may need to have someone else available to do the transfer.

Between noon and 6pm UTC from this Thursday and onwards will work on my side.

13:00 UTC should be possible. Let us work out the steps ahead so I can put in the freeze break exceptions and get them approved.

The protocol could be:

1) Something I need in advance: settings.py file (you can delete the current db password and location)
I'll need the email sending credentials.

2) put site in the read only mode (check settings -> access control -> make site read only and set an appropriate message)

3) make the dump and let me know

4) once the migration is done you'll switch the IP address which I will provide

I assume we'll want to test it in staging a bit, and the update with whatever new questions came in on the main site during that time?

In the read only mode there won't be new content, people will only be able to read existing content.

Metadata Update from @kevin:
- Issue assigned to smooge (was: kevin)

5 years ago

OK steps 1 and 3 I can currently do. Step 2 I will need to get administrative access to do or if you can do it we can do so. Where do you want the dump made available?

I can log in to askbot@fedorapeople.org, where I can pick up the dump files, but I won't be able to make the dump as I don't have access to the current server.

Ah, sorry - yes I will be able to put the site into the read only mode.

I will be available in the morning. Can we work together on IRC in #fedora-noc and can I get the ip address for the dns change

+askbot and I worked together today and got the system moved over to a new ip address. His team is working on various cleanups and other items.

Once we are clear this is good, I will decommission the old servers with backups retained.

The backups of the new site are not set up yet, I will do that by tomorrow.
Would you like the backup files be delivered somewhere in particular?

+askbot we need to have the feedback email changed to your team. It's emailing a pile of people, most of whom are not involved in ask anymore.

@askbot My apologies for not answering the backups earlier. Normally we have an rsync module set up which only allows small number of ip addresses to get to it (209.132.181.0/24). We would then do a regular rsync like

rsync -avSHP ask.fedoraproject.org:/backups/

and get the data.

@smooge, thank you, I will set up that directory and let you know.

Metadata Update from @smooge:
- Issue close_status updated to: Fixed

5 years ago

Metadata Update from @smooge:
- Issue priority set to: None (was: Waiting on External)

5 years ago

Login to comment on this ticket.

Metadata