One of the nodes at the start of the load has a high frostfs_node_engine_put_duration #281
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
3 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: TrueCloudLab/frostfs-node#281
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?
At the start of the load (8kb) in a series of launches, unexpected behavior is observed in the frostfs_node_engine_put_duration metric. Each time one of the nodes has a value 3-10 times higher than the other nodes.
As you can see on the last run, this state alternates.
Load profile: 8kb for 40 buckets, 800 threads.
Write-cache takes some time to initialize, could be it.
Oh, stop it is the start of the load, not service, right?
Yes, the run starts on clean nodes (without meta and data)
Metric
frostfs_node_engine_put_duration
was replaced withfrostfs_node_engine_request_duration_seconds
Tested on hardware, not reproduced:
There are some pikes, but it is regular values, not 3-10 time bigger.