* [dpdk-dev] LACP bond link broken chain due to the timeout
@ 2017-07-24 2:46 Zhangkun (K)
0 siblings, 0 replies; only message in thread
From: Zhangkun (K) @ 2017-07-24 2:46 UTC (permalink / raw)
To: dev, tomaszx.kulasek; +Cc: Liuyongan, Chengwentao (Vintorcheng), zhouzhengwu
Hi, all
I use dpdk LACP bond for long time large flow test, there are LACP broken chain. The dpdk log is as follow:
Bond 1: slave id 0 distributing stopped.
Bond 1: slave id 1 distributing stopped.
Through the analysis of code , LACP protocol packets are handled by eal-intr-thread thread, at the same time, the thread will also deal with the other driver interrupt event and query the NIC card state .
And each polling thread is waiting for a long time that is leading to the switch disconnection timeout.
On the generic x86 OS server, has it been considered that the eal-intr-thread thread can 't get a timely scheduling, and that would result in LACP timeout and Link broken chain?
Or has it been considered that LACP bond would be put on a single thread, not shared with others?
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2017-07-24 2:46 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-07-24 2:46 [dpdk-dev] LACP bond link broken chain due to the timeout Zhangkun (K)
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).