Oleg Sidorkin

The same thing happens with code-server 4.8.1/ node16 / FreeeBSD 13.

After looking at compiled JS, i think it's here: https://github.com/microsoft/vscode/blob/9db57e76e9f16f552203585a394e50cf0f7a84a7/src/vs/platform/environment/node/userDataPath.js#L98 However setting --user-data-dir ~/.local/share/code-server didn't help. Electron is 18.3.11

> I changed the default shell to bash or zsh and solved the problem by referring to the link below. Thank you @kawanakaiku > > [#5496 (comment)](https://github.com/coder/code-server/issues/5496#issuecomment-1240839804) Yes, this helped....

Hello. In case of replication these tables are required to have the same shards across all replicas, and their resharding is actually draining and recreating the log tables. Each store...

> Does it mean that chaos replication queue replace whole table, not just shards? Yes. > Queues for sorted chaos replicated tables mentioned because they have same design as sorted...

Number of shards is usually bound to a the number of pods, that read the queue and process data. And _usually_ queue resharding is done simultaneously with resharding consuming pods....

It is usually scripted in this way: 1) Create new queues with desired number of tablets 2) Switch producers to new queues (it may cause a write downtime that usually...