- **status**: assigned --> duplicate
- **Comment**:

According to ticket #3263, rde used broadcast for peer info message instead of 
directly send. It could reduce the frequency of error so we close this ticket.



---

** [tickets:#3258] rde: Failed to send the peer information response to an 
already down peer **

**Status:** duplicate
**Milestone:** 5.21.10
**Created:** Tue Apr 20, 2021 05:08 AM UTC by Hieu Hong Hoang
**Last Updated:** Tue Jun 01, 2021 12:52 AM UTC
**Owner:** Hieu Hong Hoang


Currently, rde receives the peer up/down event in a mailbox from mds. If the 
peer's down right after it's up, the mds could remove the peer from the 
subscription list before the rde process the peer up event in the mailbox. That 
leads to rde send the peer information response to an already down peer.

Following is the scenario after nodes rebooted from split-brain:

* SC-6 gave up election against SC-3 after sent the peer up event to SC-10:
~~~
2021-04-20 05:18:26.993 SC-6 osafdtmd[126]: NO Established contact with 'SC-10'
2021-04-20 05:18:26.994 SC-6 osafrded[151]: NO Peer up on node 0x2030f
2021-04-20 05:18:26.994 SC-6 osafrded[151]: NO Peer up on node 0x20a0f
2021-04-20 05:18:26.994 SC-6 osafrded[151]: NO Got peer info response from node 
0x2030f with role Undefined
2021-04-20 05:18:26.994 SC-6 osafrded[151]: NO RDE role set to QUIESCED
2021-04-20 05:18:26.994 SC-6 osafrded[151]: NO Giving up election against 
0x2030f with role Undefined. My role is now QUIESCED
~~~
*  SC-10 received the peer up event from SC-6:
~~~
2021-04-20 05:18:26.995 SC-10 osafrded[151]: NO Peer up on node 0x2060f
~~~
* SC-10 failed to subscribe the SC-6  rde service again and reported a send 
error :
~~~
2021-04-20 05:18:31.993 SC-10 osafrded[151]: WA Failed to send 
RDE_MSG_PEER_INFO_RESP(4) to 2060f00000097, and blocked for 4997190 us
2021-04-20 05:18:31.993 SC-10 osafrded[151]: NO Peer down on node 0x2060f
~~~


---

Sent from sourceforge.net because [email protected] is 
subscribed to https://sourceforge.net/p/opensaf/tickets/

To unsubscribe from further messages, a project admin can change settings at 
https://sourceforge.net/p/opensaf/admin/tickets/options.  Or, if this is a 
mailing list, you can unsubscribe from the mailing list.
_______________________________________________
Opensaf-tickets mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-tickets

Reply via email to