DAOS-19212 object: some improvement to avoid server overload#18609
Draft
Nasf-Fan wants to merge 1 commit into
Draft
DAOS-19212 object: some improvement to avoid server overload#18609Nasf-Fan wants to merge 1 commit into
Nasf-Fan wants to merge 1 commit into
Conversation
|
Ticket title is 'Aurora: performance jobs keep timing out with v2.8.0-rc1 and MDonSSD' |
Mainly include two fixes: 1. Make client to retry modification (RPC) if related IO handler hits -DER_TX_RESTART failure repeatedly. On server side, when IO handler repeatedly hit -DER_TX_RESTART, then it is quite possible that the -DER_TX_RESTART failure is related with server overload or some congestion caused RPC delay. Under such case, server retry with newer epoch may increase server workload/congestion. Then let's ask client to retry with some backoff delay. 2. Restrict inflight object RPC bulk transfer. Too many inflight object RPC bulk transfer many cause server overload and network congestion. This patch introduces new server environment variable "DAOS_OBJ_RPC_BULK_THD" to control the inflight object RPC bulk count. Once exceeds such threshold, server will ask client to retry. The default value for DAOS_OBJ_RPC_BULK_THD is 512 (per-target). The admin can disable such restriction via setting it as zero. Signed-off-by: Fan Yong <fan.yong@hpe.com>
Nasf-Fan
force-pushed
the
Nasf-Fan/DAOS-19212_2
branch
from
237e152 to
771e673
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Mainly include two fixes:
On server side, when IO handler repeatedly hit -DER_TX_RESTART, then it is quite possible that the -DER_TX_RESTART failure is related with server overload or some congestion caused RPC delay. Under such case, server retry with newer epoch may increase server workload/congestion. Then let's ask client to retry with some backoff delay.
Too many inflight object RPC bulk transfer many cause server overload and network congestion. This patch introduces new server environment variable "DAOS_OBJ_RPC_BULK_THD" to control the inflight object RPC bulk count. Once exceeds such threshold, server will ask client to retry. The default value for DAOS_OBJ_RPC_BULK_THD is 512 (per-target). The admin can disable such restriction via setting it as zero.
Steps for the author:
After all prior steps are complete: