--copies number
in froctfs-cli object put
works incorrectly #538
Labels
No labels
P0
P1
P2
P3
badger
frostfs-adm
frostfs-cli
frostfs-ir
frostfs-lens
frostfs-node
good first issue
triage
Infrastructure
blocked
bug
config
discussion
documentation
duplicate
enhancement
go
help wanted
internal
invalid
kludge
observability
perfomance
question
refactoring
wontfix
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: TrueCloudLab/frostfs-node#538
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Expected Behavior
While using
--copies-number
and disabling policer, we expect the number of object copies to equal the argument set via--copies-number
regardless of the placement policy.Current Behavior
While using
--copies-number
and disabling policer, the number of object copies is taken from the placement policy.Steps to Reproduce (for bugs)
/etc/frostfs/storage/conf.d/99_config.yml
:systemctl reset frostfs-storage
/etc/frostfs/s3/conf.d/99_placement_policy.yml
to a big number. Eg.:systemctl reset frostfs-s3
aws s3api create-bucket --bucket <bucket_name> --no-verify-ssl --endpoint-url https://<data0> --acl public-read-write
curl --head https://<data0> -k
frostfs-cli object put -r localhost:8080 --cid <bucket_cid> --wallet <wallet_path> --file 3<file_path> --copies-number <num>
frostfs-cli object nodes -r localhost:8080 --cid <bucket_cid> --wallet <wallet_path> --oid <oid>
The number of actual copies doesn't match the number se tvia
--copies-number
:Context
The only case it worked was when 1 was set in the placement policy file
and the same number was set via
--copies-number
while putting an object.Regression
No
Your Environment
SberCloud v1.3.0-65-echichindaeva
Alright. I tried to reproduce the problem described in the issue , with the cloud, and I got proper behavior when I restarted services on all hosts.
The problem is that policers that are still enabled and running on other hosts and they still replicate objects by given container policy.
Briefly: you should restart services on all hosts after the config is changed :)