@mberhault found this https://dev.to/acro5piano/specifying-user-and-group-in-docker-i2e
@mberhault thanks for your support!
@bladefist it sounds like cron should be fine for what you're doing (or dkron.io, if you want something highly available with a web UI) - I personally don't think that scheduling queries belong in CRDB's administration panel (for the same reason a CMS wouldn't come bundled with a web server). In UNIX world, it's costume for tools to do one thing, do them well, with separation of concerns. So I guess that you would write a shellscript wrapper that runs a query that you pass as argument, and in case of any failure, does a ROLLBACK. You can use "expect" together with a command-line sql client for that (run the query, "expect" for the right return, etc). https://linux.die.net/man/1/expect
Cool, dkron.io looks useful. I'll go down this path. Thanks!
@bladefist on your question about
IMPORT TABLE vscockroach sql < dump.sql. The import command will make use of all the nodes of the cluster to convert the csv file. The cockroach sql command will have some ability to run a few of the commands in parallel but will be much less so then the IMPORT TABLE command.
Happy to provide logs of the sydney instance if someone wants to walk me through how :)
@timaksu_twitter: is there anything pointing towards CockroachDB having issues, or just Nakama? If the latter, you may want to ask on their channel first: https://gitter.im/heroiclabs/nakama
We never have to reset Nakama
At random times we wont be able to log in to our application anymore - Nakama will still be up but cockroach will be dead, so we just have to restart it manually (dont even need to restart Nakama) for it to start working again
Anyway, do you see anything of interest at the end of the logs of the crashed nodes?
If it's a fatal error, you should see a log line starting with
F101015 ... (the date and time) followed by a giant stack trace.
the binary built on debian doesn't work on centos
looks like this issue is fixed in cockroachdb/cockroach#12694
so to make binary that works across linux distros - do I have to build it in docker image based on ubuntu xenial?
random drive-by comment: the DB selector on https://www.cockroachlabs.com/docs/stable/cockroachdb-in-comparison.html could be made more obvious, or drawn attention to
Where is this data stored now on disk?
The sql client also prints out the clusterId
Where is it reading all this data from?
For your data directory, file descriptors still function properly after a file is deleted, so any open files are still fine. However, the moment the node tries to open a file that's supposed to exist, it will crash.
So if I insert a large amount of data using the sql client, the node will try to create a new sst file in cockroach-data dir, but since the dir doesn't exist anymore, it will crash. Is my assumption correct?
What happened with the optimizer page?
https://www.cockroachlabs.com/docs/v2.1/sql-optimizer.html is the first (and only relevant) link in google, and it is a 404.
https://www.cockroachlabs.com/docs/v2.1/sql-optimizer.html is the first (and only relevant) link in google, and it is a 404.
@isoos it’s now located at https://www.cockroachlabs.com/docs/v2.1/cost-based-optimizer.html - I’ll mention that to the docs team, thanks for the heads up