GitLab GitLab 13.3 Release Product Kickoff Review, 16 Jul 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: GitLab 13.3 Kickoff - Enablement:Distribution

Description

13.3 kickoff for the Distribution team

A

Welcome to the distribution team kickoff for get lab 13.3, I'm larissa lane, I'm the product manager for distribution. We have a lot of great things planned for the 13.3 release. Today, I'm just going to cover about seven or eight of the highlights of what we'll be working on.

A

First up, we've been working really closely with red hat, to build an operator that supports running gitlab itself on openshift.

A

As part of this work, we want to investigate whether we can use the get lab helm chart and have the operator consume that helm chart so that we can continue to configure new gitlab features within the helm chart and then have the operator handle day two operations of deploying on openshift, such as backing up gitlab and upgrading getlab.

A

So we have a proof of concept scheduled for 13.3 to do some more investigation on how this might work.

A

Next up in get lab 13.2, we added support for using postgres 12 for single node, installs and single node upgrades. The next phase of the postgres 12 support will be validating that it works on multi-node installs, so in stores, where you have multiple postgres nodes, set up using replication and failover and also looking at validating geo.

A

So one thing to note here is that we have, in the past, used rep manager to replicate and failover postgres nodes.

A

The version of rep manager that we use doesn't support postgres12. So if you are interested in using postgres12 just keep in mind that you'll need to migrate from rep manager to petroni, which was newly added into the omnibus package and use petroni for handling replication and failover.

A

So you'll need to do that. Migration first before you can upgrade to postgres 12.

A

And also on the theme of petroni, while we added petroni as an experimental feature in 13.2, we are still validating using petroni with a geo secondary site configured. We did some testing of this in 13.2.

A

We ran up against some replication issues for replicating to the geosecondary, so in 13.3 we'll continue to be testing and doing more work to support using petronee with geo, and next we have been doing some work for the last couple of milestones to look at the best way to encrypt old, app secrets that are stored in gitlab configuration files. So these are the secrets that are used to connect to your ldap server for signing into gitlab using ldap credentials.

A

uh We we've been working on a proof of concept in 13.2 and we're going to continue working on that proof of concept throughout the 13.3 milestone.

A

And another item that we were working on on 13.2 made really good progress, but I haven't quite completed the work so we'll continue to work on it in 13.3 is adding support for configuring prefect in the gitlab helm charts.

A

So this is, if you want to set up multiple getaway nodes for scaling out a gilly cluster, and this will support being able to configure that in the getlab helm, charts.

A

These next two relate to being able to support kms encryption when storing objects in aws.

A

So a couple of things we're working on here that have been ongoing for a little while are this one is being able to configure credentials for workhorse to connect to aws object, storage and this next one is a really nice improvement. That's been made um sorry, the the last one, and this one both relate to being able to configure it in the gitlab helm charts.

A

um So this one is consolidating the settings for configuring object, storage. So this example here is the way that object, storage used to be configured in the helm chart.

A

You would have a block for each type of object and you would have to specify the connection details for connecting to the aws object, storage for every object, type block.

A

The new way which is shown here is that you only have to configure the connection details once and then that those connection details can be used for all the different types of objects that you want to store on aws. So a nice time saver. There.

A

Next, we have had some community work uh that is um working on supporting running, get lab itself on arm 64., so fantastic community contribution has been made there and there's a little bit of work. We need to do on our side to support that and we have that schedule for 13.3. So this is to build out um the hardware and pipelines to support building an omnibus package for running gitlab on arm64.

A

And finally, we've been making great progress on the orchestrator tool, which is the tool that we're building to automate deploying a large-scale installation of gitlab that scaled across scaled out across multiple nodes, including a geosite, sorry a geosecondary site.

A

um So we recently closed out the mvp for this work and the mvp is for internal use to use the orchestrator to spin up environments for testing gitlab upgrades on large scale deployments and with geo, and that mvp includes support for creating and configuring three console nodes, three database nodes and one application, node and also a geosecondary site.

A

So the next phase of work that we have coming up in 13.3 is to expand on the node types, so we'll be adding a load balance, support for deploying a load, balancer, kiddie nodes and redis nodes.

A

Also, a lot of other work happening for the orchestrator in 13.3 uh so feel free to jump in and use this these filters here. So if you feel to buy the uh orchestrator label and the 13.3 milestone, you'll be able to dig in a little deeper on the work, that's being done there. Thank you so much. I look forward to providing more updates in a month from now.

A