Can´t perform manual SCPU on CPP IV

[martinlopez]

I Guys, yesterday we upgraded our 81C release 25.40 to CS1000M Relese 5.50 with Cores CPP IV.
I know that is not necessary to have Switch CPUs enabled on OVLY for midnight routine, but after this upgrade I never got manual Switch CPUs by command line "SCPU" on LD 135 just I received "Graceful switchover not executed - local ELAN health is better than remote".
Please can anyone explain to me what happened?. I want to do that in order to test both CPUs and fill our daily routine as I did it with Motorola CPUs.

Thanks a lot...



########################################

.stat cpu

cp 0 16 PASS -- ENBL

SYSTEM STATE = REDUNDANT
DISK STATE = REDUNDANT
HEALTH = 26
VERSION = Feb 6 2008, 21:13:11
Side = 0, DRAM SIZE = 512 MBytes

cp 1 16 PASS -- STDBY

SYSTEM STATE = REDUNDANT
DISK STATE = REDUNDANT
HEALTH = 26
VERSION = Feb 6 2008, 21:13:11
Side = 1, DRAM SIZE = 512 MBytes

=========================================
Summary of Local System Resource (side 0)
=========================================
File Descriptors
-----------------
alloc 65
free 1983
total 2048

Unprotected Heap (bytes)
------------------------
alloc 40000088
free 208853624
total 248853712

Protected Heap (bytes)
----------------------
alloc 1347460
free 2846844
total 4194304
.****

>ld 137
CIOD000
.stat rdun
RDUN ENABLED

>ld 135
CCED000
.scpu
OK
.
SRPT018 LCS: Graceful switchover not executed - local ELAN health is better than remote.

 

[fritzthecat]

Hello MArtin,

go in load 137 and do a stat hsp, look for

.stat hsp
LCS HSP STATE is UP
HSP LINK CARRIER: OK
Auto Negotiation: Enabled
Auto Negotiation Completed: YES
Actual Line Speed: 100 Mbps
Actual Duplex Mode: Full Duplex

Auto negotiation should be enabled and duplex mode should be full

then go in load 135 and do a stat health, both sides should look the same, if there is a difference try to reboot the inaktive CPU

also look in ld 117 if both hosts are configured prt elnk

good luck

 

[SL1M1]

Maywell be a silly question - Have you got both of your CPU's physically connected to the ELAN ?

Cheers!!

 

[DFKSydney]

Remember that the HSP needs to be connected (reversal cat 5 cable) plus configured in LD 117 as well.

 

[gwebster]

With CPP of any variety, both CPUs are normally running and doing tasks but only one of them is actually taking the call processing load. This is definitely different from older processors where the standby was actually not doing anything. It is one of the reasons why the CPP CPUs initially were not configured to automatically switch. The other reason was that the switchover was not as "graceful" with CPP so the belief was not to do it unless necessary.

 

[golfdoctor]

Hi martinlopez,

Had a similar situation if I remember correctly with a previous version than 5.5. To me with RDUN showing ok , the cores are talking to each other via the ELan (could be wrong though) If the other suggestions do not work then I might suggest ensure all Signalling Servers and cores are on the same 5.5 release and are patched up to date.Since the health monitor is showing both cores the same, you should be able to switch cores.

Even though when Rlse 4.5 or so came out and it was suggested not to have the switchover in the midnight routines I have always left it in place if for no other reason to ensure that the cores actually work the way they are supposed to. Am running CPIV, CS1000M and it is in the midnight routines to switchover- Not a problem....


TGD

 

[mikeyb123]

we are experiencing the same issues with a couple of our cs1k @ r4.5 and we keep getting it. we resolve by reloading the inactive side it works for a couple of weeks then just packs up again.

most annoying.....

It's not getting any smarter out there. You have to come to terms with stupidity, and make it work for you.