4 ms·
I know its not as sexy as new features, but I really wish gitlab took a look at performance. Currently its sitting at 9GiB of Memory on my dedicated server box
by dijit 9y ago
I know its not as sexy as new features, but I really wish gitlab took a look at performance. Currently its sitting at 9GiB of Memory on my dedicated server box for basically no users. I tried it on smaller instances but it was so painfully slow I ended up getting a cheap dedicated server from Hetzner.
- sondr3 9y agoWhich is why I personally use Gitea, which is super light and performs very well. But it only has a fraction of the features of GitLab so it's not an apples to apples comparison, but my personal projects and my friends it works just fine. I combine it with Drone for testing and it totally suits my needs, and I host everything myself on a $5 VPS.
- wjkohnen 9y agoWe use GitLab on prem at work because we want the features. For private stuff or super small teams I'd chose Gitea as well. And/or GitLab.com. GitLab's performance is a major pain point for us. We throw an absurd amount of hardware at it for mid sized teams.
- nwatson 9y agoBitbucket. It's cheap.
- ausjke 9y agosame here, gitea works for probably 90% of the users IMO, I use it to host everything these days. tried gitlab a couple of times in the past, not impressed by the performance yet. have not used drone yet, will give it a shot.
- trashcan 9y agoAgreed. I really would prefer to use Gitlab, but at the moment I'm using Gogs on my Linode instance because it runs smoothly in <1GB of RAM.
- fb03 9y agoThat's what I've been using and idk if I need that much many features. Gogs gives me https+ssh, and a proper sane ui to quickly create accounts and config things. It's so lightweight I don't see me migrating to bigger stuff anytime soon. Any clue on what gitlab-only features people need that makes Gogs unusable or unfitting? make me really curious
- boudin 9y agoThere's the whole CI part. Gitea miss code comments for code review. Gitlab also have project management tools like issue boards And surely some other things i haven't looked into Gitea definitely do less but do it well and much in a much lighter way
- ptman 9y agoIt may be coming in the near future: https://github.com/go-gitea/gitea/pull/3748 https://github.com/go-gitea/gitea/pull/3748
- boudin 9y agoI switch to gogs (and now gitea) for this reason 2 years ago. It does much less in term of features though
- scottfr 9y agoThey have a dedicated section about performance improvements in the release notes. It links to a list of 53 commits from this release targeting performance: https://gitlab.com/groups/gitlab-org/-/merge_requests?scope=all&utf8=%E2%9C%93&state=merged&label_name%5B%5D=performance&milestone_title=10.7 https://gitlab.com/groups/gitlab-org/-/merge_requests?scope=...
- sytse 9y agoThanks Scott for linking to that. We're very much trying to improve the performance of GitLab. I do think that most of our performance improvements are to improve the end-user experience. So I totally understand dijit's request to reduce memory. Today we recommend 4GB of memory to run GitLab https://docs.gitlab.com/ee/install/requirements.html#memory https://docs.gitlab.com/ee/install/requirements.html#memory One big thing would be if we can run the Ruby code multithreaded https://gitlab.com/gitlab-org/gitlab-ce/issues/3592 https://gitlab.com/gitlab-org/gitlab-ce/issues/3592 You can see I created that issue 2 years ago and there has been recent activity on it. Right now GitLab the company is focussed on making complete DevOps https://about.gitlab.com/2017/10/11/from-dev-to-devops/ https://about.gitlab.com/2017/10/11/from-dev-to-devops/ happening. So any help on the multi-threading is appreciated. The first thing would be to implement https://github.com/covermymeds/rubocop-thread_safety https://github.com/covermymeds/rubocop-thread_safety
- jbergstroem 9y agoIt might be worth mentioning that they use jemalloc which greedily allocates memory. I "found out" while running into an issue and disabled it[1] which shaved a fair amount of memory from allocation. At the end of the day I guess it didn't matter - for me - seeing how I didn't need that memory for anything else. Also, the cookbook for setting up postgres also takes your total ram into account. 1: https://gitlab.com/gitlab-org/omnibus-gitlab/issues/3313 https://gitlab.com/gitlab-org/omnibus-gitlab/issues/3313
- jasone 9y agojemalloc does not greedily allocate memory, and it's not clear to me how you would think so from reading that gitlab issue discussion. I can well imagine jemalloc not behaving as desired under certain conditions, but if so it's surely more involved than wanton naivete.
- jbergstroem 9y agoI explained poorly. I ran into issues with updating jemalloc as part of a software update. While debugging the issue I disabled jemalloc (unsetting the ld_preload) and noticied a sharp difference in how much memory ruby had allocated. I guess _pre_allocating is incorrect and a better explanation is _not releasing it_. Thanks for rebutting my argument and giving me the opportunity to elaborate.
- Anthony-G 9y agoFor the past few years, I've been successfully running Gitolite[1] on a server with 1GB RAM. I recently checked out Gitlab as I thought it would be useful to have a nice web interface that allows users to easily see what projects are hosted and their corresponding URLs. I had to back away when I learned of Gitlab’s resource requirements. Based on the sibling comments here, it looks like Gogs/Gitea should do what I want. [1]: http://gitolite.com/gitolite/index.html http://gitolite.com/gitolite/index.html