Hacker News new | ask | show | jobs
by mindcrime 1485 days ago
Among other things, I'm working on migrating a bunch of infrastructure from Linode to OVH. As part of that, I'm working to ensure that every. last. drop. of this stuff has an automated deployment using Ansible. Having a bunch of hand-rolled, "pet" servers causes issues.

One very noticeable problem is that it's really easy to let things get out of date because upgrading them is painful without automation. So, for example, our CRM application is really old. Like, really, super, painfully old. Which means the other half of this project is "currency" - that is, updating software to current versions, and generally modernizing things.

The web server and xmpp server have been updated, and now I'm working on the server that hosts SugarCRM (which is no longer OSS, so we're migrating to a different product which was forked from Sugar), Mediawiki, and some locally developed stuff.

There's a few other miscellaneous improvements being made as part of all this as well: migrating all DNS records to ONE provider is one. Before, we had stuff on Route 53, and Rackspace. The latter was a legacy of our first website having been hosted on Slicehost way back in the day, and so we wound up being Rackspace customers for a time after they acquired Slicehost. We moved away from them for VPS's years ago, but DNS records for two domains got left there, and we've been dutifully paying an invoice of like $1.50 every month for DNS hosting all this time. Nothing against Rackspace, but I want things more centralized, so that is migrated to Route 53 now.

Should probably look at moving all of our domains to one registrar as well, eventually.

1 comments

heh, I recently moved a bunch of sites over to Cloudflare to avoid paying for monthly Route 53 fees for doing basically nothing