[LU-8429] Add option for gnilnd to not reconnect after connection timeout Created: 21/Jul/16  Updated: 06/Dec/16  Resolved: 14/Oct/16

Status: Resolved
Project: Lustre
Component/s: None
Affects Version/s: None
Fix Version/s: Lustre 2.9.0

Type: Bug Priority: Minor
Reporter: Chris Horn Assignee: Chris Horn
Resolution: Fixed Votes: 0
Labels: patch

Severity: 3
Rank (Obsolete): 9223372036854775807

 Description   

When routers time out a client connection during a catastrophic
network disturbance like a cabinet EPO, there still may be
traffic from the file system that is using the router for the
return path to the client. This will cause a new connection to try
to be formed before the network has quiesced causing multiple failed
connection attempts which need to be put in purgatory since they could
possibly connect in the future. This can cause the gart space to be
consumed with registrations.

So we'll add an option to not reconnect after connection timeout



 Comments   
Comment by Gerrit Updater [ 21/Jul/16 ]

Chris Horn (hornc@cray.com) uploaded a new patch: http://review.whamcloud.com/21459
Subject: LU-8429 gnilnd: Option to not reconnect after conn timeout
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: e226eff34f1c9331fa73619e97851c57d7808b09

Comment by Gerrit Updater [ 13/Oct/16 ]

Oleg Drokin (oleg.drokin@intel.com) merged in patch http://review.whamcloud.com/21459/
Subject: LU-8429 gnilnd: Option to not reconnect after conn timeout
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 99bc4ba277637656f6329a67158af6cee7070b48

Comment by Peter Jones [ 14/Oct/16 ]

Landed for 2.9

Generated at Sat Feb 10 02:17:27 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.