Who is using rethinkdb

MoOmer on Oct 9, prev [—]. I'm torn between supporting open source businesses, and calling this and almost all of your previous submissions spam. TheMissingPiece on Oct 9, parent next [—]. Definitely not spam. I always hear people ask who is using RethinkDB in production, so just thought I'd help answer that here. I get where you're coming from, though :. I greatly appreciate this post. I would love to hear about other people who are using RethinkDB.

It is an interesting and unique tool, and fits certain use cases very well. Rethinkdb was seeded by ycombinator and Andreessen Horowitz. There is no way these posts would ever get removed even if most people did feel it was "spam". Dang is incorruptible which is why YC put him in charge of HN. If enough people flag something as spam or he felt something is spam , the post would removed faster than we could blink.

Isn't he being paid by YC though? And there are frequently special hiring threads for YC companies. Anyway there's a very thin line between spam and not spam, and is generally based on whatever people want to see at the time.

TheMissingPiece on Oct 10, root parent next [—]. Next, in early January, I started the process of writing code to migrate all the data from RethinkDB to PostgreSQL, with minimal downtime so one big migration, then incremental updates.

I thought this would take a few hours, but it ended up taking nearly a month! I have a lot of data — one table had million records in it… Another obstruction is that PostgreSQL is statically typed, whereas RethinkDB is very much not… and this exposed tons of subtle issues in my data.

In addition, with PostgreSQL it was obvious and trivial to impose conditions on my data, e. By this point, I knew both query languages pretty well. PostgreSQL was faster, usually 5x faster, sometimes only 2x faster, and often even 10x faster.

I am not providing my data or one single proof of my claims about speed. Again, this is watercooler talk. I have a couple hours to share my experiences with the world, and then I have to get back to work.

Backups, which involve dumping full tables from the database, were an order of magnitude faster with PostgreSQL. Often the first Python implementation of an algorithm is nice and illustrative and works; then you re-implement it in Cython, change algorithms, etc. I imagine databases are similar. Using 10x more disk space means 10x more reading and writing to disk, and disk is way more than 10x slower than RAM….

I spent a huge amount of time worry about connection pooling with RethinkDB to get better concurrency, finally just writing my own.

By optimizing everything, the load on the database and the web servers is now overall very low, and can easily be handled over a single connection. There simply is no need for a connection pool for my application, since PostgreSQL is so fast. After all the awesome microbenchmarking above, I expected that when we went live it would be way more efficient than RethinkDB.

On a nervous quiet Saturday morning, we switch the live production site over, and everything looked reasonably good for a while. It made no sense. I made the database server faster and spun up way more web servers, which basically worked… but seemed weird. I panicked for a while, mulled over the problem, and kept raising the number of web servers, etc. This sucked. After convincing myself not to give up and shut down SMC for good, I calmed down and studied a lot of logs and found a PostgreSQL query that was taking 15s sometimes and locking the other queries.

I then tried an instance of this query directly in psql, and it took only a few milliseconds. I learned how to impact the query planner, and then this query went back to taking only a few milliseconds for any input. With this one change to influence the query planner to actually always use an index I had properly made , things became dramatically faster. The Node. This was another observation I made by looking at log files. Basically it is everything that starts with postgres- here.

SageMath, Inc. Thanks to your contributions to Rebirth DB, The Linux Foundation generously allowed us to merge the project back so we are one community again. As KittyBot mentioned, we were able to build the RethinkDB website again and set up continuous deployments for it, which allows us to publish new blog posts, configure the stripe account, and of course, update the content which is a bit outdated.

Also, I'm one of the maintainers of the Python client. Nowadays, I had busy days and I had not too much time contributing to the project but it will change. At the moment with KittyBot we are collecting every blocker and obstacle which need to be eliminated, and we are setting up both short and long term plans to keep the community alive and loop more guys into the development of the DB.

Jsieler , I'm glad that you created the discord channel for RethinkDB, but at this point, I would like to note that we got back the access to Slack. To set up a self-invite flow for Slack is also on our roadmap so we will be able to keep the community at one place.

The communication was not our strength with the future of the project. We noted this and this will change in the future. I would like to thank you for those who contributed to the project since it was open sourced and for those who stayed with us. So RethinkDB will continue to be the same.

No, I'd think that would be post RethinkDB will be unchanged. I am concentrating on one. The advantages I'll put my marketing copy elsewhere. Many thanks for the update, srh. I'm sorry to hear about the stress - personal wellbeing is paramount.

It might be worth a shot before going down the licensing path. It's preferable, in principle, for the project to be open source and available for all to use at no cost, if viable. It could also help to revive RethinkDB as a whole, by being accessible to newcomers and returning users, no questions asked.

Crowdfunding could be a little less pressuring than direct funding, given that it is common for crowdfunded projects to take some time to come to fruition, and people know that. It'd require some organisation to get off the ground, including spreading the message among the remaining Rethink community. Perhaps KittyBot and others could lend a hand with this? If it failed to raise enough funding to support the project, the funds could be refunded, and the project could fall back on the licencing approach.

But, as I say, it seems to be worth a shot, for the chance of keeping the flame of open source and open access alive.

I was looking at FoundationDB but it was too confusing to setup and use. There was a time I was running a cluster across my phone and laptop using zerotier and termux. That was nice. I've also looked into Foundation and found its administration a little daunting.

Unfortunately, I don't recall the precise threads. I think the takeaway was that, while Foundation might not be as straightforward to administrate as Rethink, ReQL-on-FoundationDB would be ultimately workable, with the benefits in performance, stability, and features making it worthwhile.

Yeah me too. I don't want to move away from RethinkDB. KittyBot gabor-boros Can I suggest that now you can deploy the website that you add a banner to the home page stating some to the effect that the project is not dead and is actively being resurrected? I use the website daily and I had no idea this conversation had been started until srh answered one of my old posts and pointe me to it.

It would encourage use. On the Discord front - yet another channel? With Slack access back I don't see why we wouldn't revert to that. Spectrum is pretty clunky IMO. AaronRobinson Thank you for your suggestion. Last week I spent some time to be able to redeploy the site again on an automated basis. Currently, with KittyBot , we are revisiting the content and the website related GitHub issues. My goal for the next week is to redeploy the site. This does not mean that everything will be up-to-date by then, but we will be able to deploy changes.

As I mentioned before, communication was not our strength until now. I would like to inform you about the changes around RethinkDB on a weekly basis at least I would try it. Related to Slack, I'm actively working on to open up slack again for everyone who would like to join. I think this will be done by the middle of next week.

WoLfulus I totally understand this. Although I'm not against discord really , we have a lot of people at Slack, and we need to get their confidence back. If we switch platform now for communication, I believe, we miss the opportunity to reach those members.

Please consider many of the main rethinkdb users Jumping between KV Stores are showing the lack of direction too.. I'll pin this to the top of the channel, too. Here's the tweet, too, if you'd like to share. We are trying to collect everything before the next release, and this is on my list as well, to at least mention the project on the website.

Personally, I would like to have bigger support for cloud-native tech stacks. What do you mean you're trying to collect everything?

What needs to be done before the next release?