Deuxfleurs/garage
41
56
Fork 73
Code Issues 99 Pull requests 14 Releases 30 Wiki Activity

Fix unbounded buffering when one node has slower network #792

Merged
lx merged 2 commits from fix-buffering into main 2024-03-28 12:40:28 +00:00
Conversation 1 Commits 2 Files changed 11 +228 -58
lx commented 2024-03-27 15:07:43 +00:00
Owner
Copy link

  • Add mechanism to limit the number of blocks buffered in RAM as they await being sent to storage nodes. Backpressure is applied to the clients when the limit is reached. The size of this buffer can be controled using the config parameter block_ram_buffer_max

  • Add metric block_ram_buffer_free_kb to visualize memory still available for buffers (when it drops to zero, backpressure is applied)

  • Change the network scheduling algorithm to send requests one after the other, instead of least-attained-service that was equivalent to a round-robin and would cause all blocks to be buffered simultaneously

Related: #788

- [x] Add mechanism to limit the number of blocks buffered in RAM as they await being sent to storage nodes. Backpressure is applied to the clients when the limit is reached. The size of this buffer can be controled using the config parameter `block_ram_buffer_max` - [x] Add metric `block_ram_buffer_free_kb` to visualize memory still available for buffers (when it drops to zero, backpressure is applied) - [x] Change the network scheduling algorithm to send requests one after the other, instead of least-attained-service that was equivalent to a round-robin and would cause all blocks to be buffered simultaneously Related: #788
lx changed title from Fix buffering issues (fix #788) to Fix unbounded buffering when one node has slower network (fix #788) 2024-03-27 15:09:19 +00:00
lx force-pushed fix-buffering from 9613f9495e to 85f580cbde 2024-03-27 15:23:01 +00:00 Compare
lx commented 2024-03-28 12:02:36 +00:00
Author
Owner
Copy link

Confirmed that this avoids unbounded memory usage growth inducing OOM kill on a test deployment when sending a single 8GB file to an S3 API server on localhost, which had to send to two other storage nodes, one a bit slower than the other.

With the default value of block_ram_buffer_free_kb (256MB), the RAM usage of Garage as measured by RssAnon is stable under 300MB during the transfer, and the overall speed of the transfer is correctly capped to the speed of the slowest internet link between nodes. Without the patch the API server received data faster than it could send back to storage nodes and RAM usage grew to 1GB, 2GB, 3GB, ... and provoked an OOM kill of the daemon.

Confirmed that this avoids unbounded memory usage growth inducing OOM kill on a test deployment when sending a single 8GB file to an S3 API server on localhost, which had to send to two other storage nodes, one a bit slower than the other. With the default value of `block_ram_buffer_free_kb` (256MB), the RAM usage of Garage as measured by RssAnon is stable under 300MB during the transfer, and the overall speed of the transfer is correctly capped to the speed of the slowest internet link between nodes. Without the patch the API server received data faster than it could send back to storage nodes and RAM usage grew to 1GB, 2GB, 3GB, ... and provoked an OOM kill of the daemon.
lx changed title from Fix unbounded buffering when one node has slower network (fix #788) to Fix unbounded buffering when one node has slower network 2024-03-28 12:05:05 +00:00
lx changed title from Fix unbounded buffering when one node has slower network to Fix unbounded buffering when one node has slower network (fix #788) 2024-03-28 12:37:30 +00:00
lx changed title from Fix unbounded buffering when one node has slower network (fix #788) to Fix unbounded buffering when one node has slower network 2024-03-28 12:39:41 +00:00
lx merged commit ecf641d88c into main 2024-03-28 12:40:28 +00:00
lx deleted branch fix-buffering 2024-03-28 12:40:28 +00:00
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels   
action
check-aws
The behaviour should be checked against AWS S3

  
action
discussion-needed
We need to discuss if the feature is desirable and/or better specify the envisioned behavior or technical implementation

  
action
for-external-contributors
Core team does not plan working on this in the near future however an external contribution will be reviewed & merged

  
action
for-newcomers
An issue that should be newcomer friendly.

  
action
more-info-needed
More information is needed to investigate this issue

  
action
need-funding
This issue represents lots of work, we need to find money to pay our core team work

  
action
triage-required
It's not very clear what we want to do with this issue yet, more thoughts is required

  
kind
correctness
Issues of theoretical correctness that may impact edge-case scenarios

  
kind
ideas
For abstract ideas/proposal to make Garage better, to better communicate our not yet achieved aims

  
kind
improvement
Feature requests that do not fall in other categories

  
kind
performance
Deceptive performance observed OR idea to improve performances

  
kind
testing
Functional tests, integration tests, Jepsen, etc.

  
kind
usability
Propositions to make Garage easier to operate, by improving error reporting, CLI, etc.

  
kind
wrong-behavior
Unintended behavior

  
prio
critical
Could impact integrity or availability of production clusters

  
prio
low
For issues that we don't want to spend time on until made absolutely necessary

  
scope
admin-api
Issues concerning the administration API

  
scope
background-healing
Background healing covers repair tasks, anti-entropy, merkle trees, etc. ; everything related to replicating data appropriately.

  
scope
build
Issues with Nix, Drone, Rust Compiler, etc.

  
scope
documentation

  
scope
k8s
Kubernetes related, including Helm Charts

  
scope
layout
Linked to the cluster layout and storage management

  
scope
metadata
Metadata Engine (LMDB, sqlite, etc.)

  
scope
ops
Operations related to operations/maintenance issue, tooling related (like CLI missing info ; impossible, too dangerous, too complicated migration, etc.)

  
scope
rpc
Linked to RPC

  
scope
s3-api
Related to the S3 API (especially compatibility issues).

  
scope
security
Cryptography-related, auth-related, authorization-related issues

  
scope
telemetry
OpenTelemetry (traces, metrics, etc.)

No labels
action
check-aws
action
discussion-needed
action
for-external-contributors
action
for-newcomers
action
more-info-needed
action
need-funding
action
triage-required
kind
correctness
kind
ideas
kind
improvement
kind
performance
kind
testing
kind
usability
kind
wrong-behavior
prio
critical
prio
low
scope
admin-api
scope
background-healing
scope
build
scope
documentation
scope
k8s
scope
layout
scope
metadata
scope
ops
scope
rpc
scope
s3-api
scope
security
scope
telemetry
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: Deuxfleurs/garage#792
No description provided.
Delete branch "fix-buffering"

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?