I've been looking at distributed #sql #databases for a while, like #cockroachdb and #yugabytedb. Does anyone have any experience running these on prod? What are the foot guns one might encounter while running these databases ?
#sql #databases #cockroachdb #yugabytedb
Distirbuted #keycloak with #cockroachdb
https://dantheengineer.com/keycloak-on-distributed-sql-cockroach-part-1-2/
We had a great fireside chat with our friends at CockroachDB -
- Why does #opensource matter
- Potential collaborations between #CockroachDB & #FerretDB
- The future of the database community
Read the entire blog post here --Â https://blog.ferretdb.io/community-matters-fireside-chat-cockroachdb/
#opensource #cockroachdb #FerretDB
Tell me you want to vendor lock-in your customers without telling me you to vendor lock-in your customers
https://github.com/cockroachdb/docs/issues/3540#issuecomment-483395275 #cockroachdb
https://cacm.acm.org/magazines/2023/6/273229-foundationdb-a-distributed-key-value-store/fulltext — strangely similar to the internals of #dynamodb and invented at about the same time.
FoundationDB is the “metastore” used on Snowflake. It is also used internally at Apple in projects they don’t disclose.. . it sounds like a cool system, too hyperscale for me to use right now..
In a universe where #cockroachdb cross-breeds w/ foundationDB we would get a really amazing open source distributed DB.
If you are using #CockroachDB, integrating with PolyScale.ai for caching and pooling will lower latency, increase query performance, and help you easily scale to dozens of global regions. Read the #DZone blog by Artem Ervits here: https://dzone.com/articles/reducing-network-latency-and-improving-read-perfor
@zaythelegend @Sknashville @andisheh @et @ShmosKnows @alexip718 @maggie_a_lee
Problem solved! I was excluding #MARTA buses marked as delayed and was only including buses marked as canceled. The fix was less complex than I anticipated, so no fancy AI needed (yet).
I'll move forward on getting the public data endpoint online - about 25% done with that.
I'll have a public bucket with static JSON output, but the actual data will sit in #CockroachDB for future live-query purposes.
1/
Reading '#CockroachDB: The Definitive Guide' liking the idea of a horizontally-scalable, Postgres client/wire-compatible #DB wondering what are the cons? When would #Postgres be the better option?
#cockroachdb is interesting... it appears that its backup statement does not expect or support a local destination
Today:
* Bullet proof the homelab #ansible provisioning
* Install #LongHorn on my #k3s cluster
* Get #cockroachdb installed on cluster
#ansible #Longhorn #k3s #cockroachdb
What happens if you put #MariaDB #Xpand against #CockroachDB? đź‘€
We predict #MariaDB #Xpand will crush #CockroachDB. đź‘Š
Let’s pull up the performance results: https://bit.ly/3Nq0U5X
Drei Fragen und Antworten: endlich echte Innovationen bei Cloud-Datenbanken
Mit dem Wechsel in die Cloud stehen erstmals seit langem komplett neu gedachte Datenbanken zur VerfĂĽgung. Wir werfen einen Blick auf Yugabyte & Co.
#AmazonWebServices #CloudComputing #CockroachDB #Datenbanken #GoogleCloud #Kubernetes #LinuxundOpenSource #MicrosoftAzure #SQL #Vitess #YugabyteDB #iX #News
#amazonwebservices #cloudcomputing #cockroachdb #datenbanken #googlecloud #kubernetes #linuxundopensource #microsoftazure #sql #vitess #yugabytedb #ix #news
I also had a chance to test #cockroachdb in practice. I wanted to get rid of separate #postgresql instances for each service and deploy a single DB cluster. Recommended minimum of RAM for crdb is 2Gi, which is of course too much, while a pgsql instance with minimum load runs OK on 256Mi. You can deploy a crdb cluster with 384Mi on each node, but you will suffer from node crashes for even simple management actions, so it's a no-go too. My current plan is to deploy a HA pgsql cluster instead.
I also had a chance to test #cockroachdb in practice. I wanted to get rid of separate #postgresql instances for each service and deploy a single DB cluster. Recommended minimum of RAM for crdb is 2Gi, which is of course too much, while a pgsql instance with minimum load runs OK on 256Mi. You can deploy a crdb cluster with 384Mi on each node, but you will suffer from node crashes for even simple management actions, so it's a no-go too. My current plan is to deploy a HA pgsql cluster instead.
So, auto migrations with #rust sqlx on #CockroachDB? Just disable locking and you're good #yolo :P
So, auto migrations with #rust sqlx on #CockroachDB? Just disable locking and you're good #yolo :P