pool: Immediately mark maintenance node as unhealthy #283

New issue

Closed

opened 2024-10-11 10:07:15 +00:00 by alexvanin · 0 comments

alexvanin commented

2024-10-11 10:07:15 +00:00

Owner

Expected Behavior

When active client receives maintenance mode error from object service, immediately switch to another healthy client in next request.

Current Behavior

Maintenance mode error considered as any other network error and increments threshold counter by one.

Possible Solution

Increment threshold variable to a threshold limit value, so it overflows the limit.

Steps to Reproduce (for bugs)

Put node in a maintenance mode
Send object.Put request with SDK Pool component
Send one more object.Put request

Context

This is similar to #278
Maintenance mode error is not an accident error, so we can rely on this status and switch client.
Otherwise we are having performance penalty when SDK Client has to decline multiple requests until threshold value has been reached or rebalance interval has been passed.

Regression

Your Environment

SDK Pool from frostfs-s3-gw v0.31.0-rc.2 (1b67ab9608)

## Expected Behavior When active client receives maintenance mode error from object service, immediately switch to another healthy client in next request. ## Current Behavior Maintenance mode error considered as any other network error and increments threshold counter by one. ## Possible Solution Increment threshold variable to a threshold limit value, so it overflows the limit. ## Steps to Reproduce (for bugs) 1. Put node in a maintenance mode 2. Send object.Put request with SDK Pool component 3. Send one more object.Put request ## Context This is similar to #278 Maintenance mode error is not an accident error, so we can rely on this status and switch client. Otherwise we are having performance penalty when SDK Client has to decline multiple requests until threshold value has been reached or rebalance interval has been passed. ## Regression No ## Your Environment SDK Pool from frostfs-s3-gw v0.31.0-rc.2 (1b67ab9608)