- **Milestone**: 4.7.2 --> 5.0.2


---

** [tickets:#1942] amf: issue with comp-failover recovery during su lock in 
N-Way model.**

**Status:** accepted
**Milestone:** 5.0.2
**Created:** Tue Aug 09, 2016 06:56 AM UTC by Praveen
**Last Updated:** Wed Aug 10, 2016 12:44 PM UTC
**Owner:** Praveen
**Attachments:**

- 
[compfail.tgz](https://sourceforge.net/p/opensaf/tickets/1942/attachment/compfail.tgz)
 (52.1 kB; application/x-compressed)


Conf: Nway model, comp-failover recovery.
steps to reprodue:
1)Bring attached configuration up.
2)Issue lock operation on SU and make sure one comp faults with comp-failover 
recovery.
3)One of the SIs will remain with standby assignments without active. AMFD does 
not perform failover/switchover of SI.

AMFD sends deletion of assignments during recivery request:
Aug  9 11:51:31.142846 osafamfd [4373:sgproc.cc:2340] >> avd_sg_su_si_del_snd: 
'safSu=SU1,safSg=AmfDemo,safApp=AmfDemo1'
Aug  9 11:51:31.142854 osafamfd [4373:mbcsv_api.c:0773] >> 
mbcsv_process_snd_ckpt_request: Sending checkpoint data to all STANDBY peers, 
as per the send-type specified
Aug  9 11:51:31.142860 osafamfd [4373:mbcsv_api.c:0803] TR svc_id:10, 
pwe_hdl:65537
Aug  9 11:51:31.142865 osafamfd [4373:mbcsv_util.c:0343] >> mbcsv_send_ckpt_dat

When AMFD gets response from AMFND in realign state, it does not faiolver the 
assignments and deletes SUSIs in SU:
Aug  9 11:51:31.160480 osafamfd [4373:sgproc.cc:1051] >> avd_su_si_assign_evh: 
id:106, node:2010f, act:4, 'safSu=SU1,safSg=AmfDemo,safApp=AmfDemo1', '', ha:3, 
err:1, single:0
Aug  9 11:51:31.160703 osafamfd [4373:sg_nway_fsm.cc:2182] >> 
avd_sg_nway_susi_succ_sg_realign:  action:4 state:3
Aug  9 11:51:31.160709 osafamfd [4373:su.cc:2047] >> delete_all_susis: 
'safSu=SU1,safSg=AmfDemo,safApp=AmfDemo1'
Aug  9 11:51:31.160714 osafamfd [4373:csi.cc:1198] >> avd_compcsi_delete
Aug  9 11:51:31.160720 osafamfd [4373:pg.cc:0165] >> avd_pg_compcsi_chg_prc
Aug  9 11:51:31.160725 osafamfd [4373:pg.cc:0175] << avd_pg_compcsi_chg_prc: 1

Because of this one SI remain in standby state without any active. 


Assignments before lock operation:
safSISU=safSu=SU1\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo,safApp=AmfDemo1
        saAmfSISUHAState=ACTIVE(1)
safSISU=safSu=SU2\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo,safApp=AmfDemo1
        saAmfSISUHAState=STANDBY(2)
safSISU=safSu=SU3\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo,safApp=AmfDemo1
        saAmfSISUHAState=STANDBY(2)
safSISU=safSu=SU2\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo2,safApp=AmfDemo1
        saAmfSISUHAState=ACTIVE(1)
safSISU=safSu=SU1\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo2,safApp=AmfDemo1
        saAmfSISUHAState=STANDBY(2)
safSISU=safSu=SU1\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo1,safApp=AmfDemo1
        saAmfSISUHAState=ACTIVE(1)
        
Assginemnts after lock operation:
safSISU=safSu=SU2\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo,safApp=AmfDemo1
        saAmfSISUHAState=STANDBY(2)
safSISU=safSu=SU3\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo,safApp=AmfDemo1
        saAmfSISUHAState=STANDBY(2)
safSISU=safSu=SU2\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo2,safApp=AmfDemo1
        saAmfSISUHAState=ACTIVE(1)
safSISU=safSu=SU2\,safSg=AmfDemo\,safApp=AmfDemo1,safSi=AmfDemo1,safApp=AmfDemo1
        saAmfSISUHAState=ACTIVE(1)
After lock operation and fault safSi=AmfDemo,safApp=AmfDemo1 remain with 
standby assignments only. AMFD can make SU3 active for this SI as it is allowed 
by the configuration.

AMFD is not performing any switchover/failover in realign state for SU level 
delete  in avd_sg_nway_susi_succ_sg_realign() function. Case of single SUSI 
delete is handled. 
Also when operation is in progress, SG was marked stable even though comp in 
locked SU is still having quiesced assignments. In Nway model, AMFD some times 
handle events in like stable state, but it should not mark it stable as it will 
open window for other admin op to progress.

Issue will be applicable to lock operation on Node also as it uses same code 
path.

Attachment includes conf and traces.


---

Sent from sourceforge.net because opensaf-tickets@lists.sourceforge.net 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
Opensaf-tickets@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensaf-tickets

Reply via email to