-
У клиента есть ssd-io диск.
Он на него пишет и в этот момент хост, на котором находится NBS рестартует.
```
-- Reboot --
Jun 01 12:39:32 : Started Network Block Storage.
```
После ребура, кл…
-
After the devices have been taken for maintenance, their statuses in DR change to "error" (due to registration of the agent without them).
Statuses of the devices should be restore to "online" after …
-
Для того, что бы ноды захватывали конфигурации из CMS их нужно переключить на новый CMS, а так же включить Configs Dispatcher, что бы захватывать изменения в CMS без рестартов сервиса.
1. Нужно пер…
-
To get rid of secrets, we've removed the playground notebooks. @ChetanChauhanSims , I think would be good to take some of the old ones and clean them up a bit? Make sure secrets are handled correctly…
-
http://tv-vod.nbs.cn/vod_storage/vol1/2024/06/06/66615d8489137/66615d848913767849.mp4
-
При рестарте DA нам нужно: пойти в DR_proxy, который идет в HIVE_proxy для того, что бы получить DR Tablet ID.
Проблема в том, что HIVE может быть не доступен какое то время.
![image](https://github…
-
In the common usecases clients don't have too many files per FS but they often do a lot of ls and stat calls. We can easily optimize such requests by caching node metadata in memory to avoid LSM looku…
-
We can implement default encryption for STORAGE_MEDIA_SSD_{NONREPLICATED,MIRROR2,MIRROR3} both at rest and in transit by encrypting data at the endpoint level - just like user key-based encryption. Th…
-
The following sensors `{project="nbs", cluster="mycluster", service="service", host="cluster", sensor="*ompressedBytesWritten", type="ssd"}` observed on our clusters show that the data that our users …
-
rename containers in TCheckpointStore to reasonable ones (checkpointsWithData/WithoutData, ...)
@leftmain has details