Fibre Channel (SAN)

Reply
Occasional Contributor
Posts: 7
Registered: ‎10-28-2009

I need ideas why FICON fails to configure online on DCX-4S

We just installed a FICON Director Brocade DCX-4S Switch with 2 16-port blades with 4Gig SFPs installed. Everything seems to be working and we have it configured for FICON only using Brocade FICON documentation. However, all our FICON4 Express CHPIDs that are connected into the DCX fail to come online. From the IBM 2096-S07 SE, I get “Invalid Attachment Failure”. From z/OS D M=CHP(01) shows the following:

D M=CHP(01)

IEE174I 08.13.51 DISPLAY M 803

CHPID 01:  TYPE=1D, DESC=FICON INCOMPLETE, ONLINE

DEVICE STATUS FOR CHANNEL PATH 01

    0  1  2  3  4  5  6  7  8  9  A  B  C  D  E  F

00C .  .  .  .  .  .  .  .  .  .  .  .  $@ .  .  .

00E .  .  .  .  .  .  .  .  .  .  .  $@ .  .  .  .

04A $@ $@ $@ $@ .  .  .  .  .  .  .  .  .  .  .  .

04E $@ $@ $@ $@ $@ $@ .  .  .  .  .  . .  .  .  .

SWITCH DEVICE NUMBER = NONE

ATTACHED ND = NOT KNOWN

PHYSICAL CHANNEL ID = 0101

************************ SYMBOL EXPLANATIONS ************************

+ ONLINE    @ PATH NOT VALIDATED   - OFFLINE    . DOES NOT EXIST

* PHYSICALLY ONLINE   $ PATH NOT OPERATIONAL

From the DCX I see

Authentication: None

portDisableReason: None

portCFlags: 0x1

portFlags: 0x1   PRESENT U_PORT

portType:  17.0

portState: 2    Offline

portPhys:  6    In_Sync

portScn:   2    Offline

port generation number:    8

portId:    cc0400

portIfId:    43120014

portWwn:   20:04:00:05:1e:b6:e1:80

portWwn of device(s) connected:

Distance:  normal

portSpeed: 4Gbps

LE domain: 0

FC Fastwrite: OFF

Interrupts:        0          Link_failure: 819        Frjt:         0

Unknown:           0          Loss_of_sync: 22950      Fbsy:         0

Lli:               246989     Loss_of_sig:  20

Proc_rqrd:         0          Protocol_err: 0

Timed_out:         0          Invalid_word: 2707420162

Rx_flushed:        0          Invalid_crc:  0

Tx_unavail:        0          Delim_err:    0

Free_buffer:       0          Address_err:  0

Overrun:           0          Lr_in:        0

Suspended:         0          Lr_out:       0

Parity_err:        0          Ols_in:       0

2_parity_err:      0          Ols_out:      0

CMI_bus_err:       0

This CHPID is defined as follows:

CHPID PATH=(CSS(0),01),SHARED,                          *

PARTITION=((CONPROD,CONTEST,PEPPROD,PEPTEST,VMSERVER,VMT*

      ESTSV),(=)),SWITCH=CC,PCHID=101,TYPE=FC

The Switch as follows:

CNTLUNIT CUNUMBR=00CC,PATH=((CSS(0),01,10)),           *

UNITADD=((00,001)),LINK=((CSS(0),CCFE,CCFE)),UNIT=2032

IODEVICE ADDRESS=0CC,UNITADD=00,CUNUMBR=(00CC),STADET=Y,*

      UNIT=2032

I am at wits end why these FICON Express 4 are not coming online. Does anyone have any suggestions? Thanks.

Peter



External Moderator
Posts: 4,985
Registered: ‎02-23-2004

Re: I need ideas why FICON fails to configure online on DCX-4S

Peter,

I'm sure NOT the FICON Specialist, but this error say to me the Port typ is wrong set.

TechHelp24
Occasional Contributor
Posts: 7
Registered: ‎10-28-2009

Re: I need ideas why FICON fails to configure online on DCX-4S

Thanks for such a quick response. How can I verify port type and what should it be for FICON?

Peter



External Moderator
Posts: 4,985
Registered: ‎02-23-2004

Re: I need ideas why FICON fails to configure online on DCX-4S

Peter,

according the Output you post here, is the Port current set as U_Port.

And according I'm not a FICON Specialist, Ports for Ficon must be set as N_Port / F_Port.

TechHelp24
Occasional Contributor
Posts: 7
Registered: ‎10-28-2009

Re: I need ideas why FICON fails to configure online on DCX-4S

Thanks again. I do not think we are even making it that far. I think the initial negotiation is failing between the FICON Express4 and the port on the DCX. I just do not know if it is a config problem on the DCX or the FICON config.

Peter



External Moderator
Posts: 4,985
Registered: ‎02-23-2004

Re: I need ideas why FICON fails to configure online on DCX-4S

Peter,

--->>> I do not think we are even making it that far.

is the Fabric correct configured ?

I found this info here, can you check if all this value are correct set in the DCX ?

Fabric Not Configured Properly

Make sure that the insistent domain ID is set and the domain IDs are unique for all switches in the fabric.

Make sure that the same binding method is used on all switches in the fabric.

Scroll through the node list and make sure there are no invalid attachments.

Make sure that the re-routing delay is turned off.

TechHelp24
Occasional Contributor
Posts: 7
Registered: ‎10-28-2009

Re: I need ideas why FICON fails to configure online on DCX-4S

It looks like we are still having CP issues on the Brocade DCX-4S Switch. I see the active CP changing back and forth and I lose connectivity. I also kept losing configuration data everytime it switched. So I never had a full configuration. I needed to do a manual HASYNC as I configure. When the CP switched, I still had my previous configuration and started from there until I was done. I then IMLed the CHPIDs from the IBM 2096-S07 SE and all was fine. I was able to vary the CUP online and was able to build enough of a configuration where I successfully downloaded the switch matrix configuration from HCD and activated it. I also successfully tested SA I/O Operations. I am getting some strange messages from SA I/O Operations that I will address with the SA folks. Right now I am thinking we are looking a lot better then this morning.

Is there any diagnostic information I can gather to address these CP switchings?

Also, our firmware is Fabric OS:  v6.2.0c . I know v6.3 is available and I was going to download and install it. Does anyone have any experience or recommendations with v6.3?

Peter



External Moderator
Posts: 4,985
Registered: ‎02-23-2004

Re: I need ideas why FICON fails to configure online on DCX-4S

Peter,

--->>> It looks like we are still having CP issues on the Brocade DCX-4S Switch.

This is sure not the reason that the FICON config fails, and i don't think you have CP problem whit the DCX.

I think the problem is in the switch config, probable is this caused by any FabOS Defect.

Can you collect please the output from "errshow" and post here ?

TechHelp24
Occasional Contributor
Posts: 7
Registered: ‎10-28-2009

Re: I need ideas why FICON fails to configure online on DCX-4S

I apologize for taking so long to respond as I have been working other issues. I do not want to leave this issue hanging as it the information might serve useful to someone else in the future. I sent the supportsave output to our vendor for analysis. I have not received any response. We are running the following levels:

Kernel:     2.6.14.2

Fabric OS:  v6.2.0c

Made on:    Mon Feb 23 19:32:17 2009

Flash:      Thu Sep 24 16:39:34 2009

BootProm:   1.0.14

Here is what we were seeing:

Time Level     Message   Service   Number    Count     Message ID    Switch    

Wed Nov 04 2009 15:25:19 EST Warning   Trace dump available (Slot 4)! (reason: PANIC)    Chassis   113  1    TRCE-1001 CCFICON1 

Wed Nov 04 2009 15:25:19 EST Warning   Trace dump (Slot 4) was not transferred because trace auto-FTP disabled.    Chassis   114  1    TRCE-1004 CCFICON1 

Wed Nov 04 2009 15:25:19 EST Warning   Trace dump available (Slot 4)! (reason: PANIC)    Chassis   115  1    TRCE-1001 CCFICON1 

Wed Nov 04 2009 15:25:19 EST Warning   Trace dump (Slot 4) was not transferred because trace auto-FTP disabled.    Chassis   116  1    TRCE-1004 CCFICON1 

Wed Nov 04 2009 15:21:49 EST Warning   Switch status changed from HEALTHY to MARGINAL.  Switch    106  1    FW-1424   CCFICON1 

Wed Nov 04 2009 15:21:49 EST Warning   Switch status change contributing factor CP: CP non-redundant (Slot4/CP0) faulty. Switch    107  1    FW-1433     CCFICON1 

Wed Nov 04 2009 15:21:14 EST Error     CP in Slot 4 set to faulty because CP ERROR asserted. Chassis   105  1    EM-1033   CCFICON1 

Wed Nov 04 2009 14:52:58 EST Warning   Trace dump available (Slot 5)! (reason: PANIC)    Chassis   102  1    TRCE-1001 CCFICON1 

Wed Nov 04 2009 14:52:58 EST Warning   Trace dump (Slot 5) was not transferred because trace auto-FTP disabled.    Chassis   103  1    TRCE-1004 CCFICON1 

Wed Nov 04 2009 14:48:42 EST Warning   Trace dump available (Slot 5)! (reason: FFDC) Chassis   98   1    TRCE-1001 CCFICON1 

Wed Nov 04 2009 14:48:42 EST Warning   Trace dump (Slot 5) was not transferred because trace auto-FTP disabled.    Chassis   99   1    TRCE-1004 CCFICON1 

Wed Nov 04 2009 14:48:24 EST Warning   Detected termination of process ficud:3345    Chassis   96   1    KSWD-1002 CCFICON1 

Wed Nov 04 2009 13:19:04 EST Warning   Switch status changed from HEALTHY to MARGINAL.  Switch    78   1    FW-1424   CCFICON1 

Wed Nov 04 2009 13:19:04 EST Warning   Switch status change contributing factor Switch offline.  Switch    79   1    FW-1439   CCFICON1 

Wed Nov 04 2009 12:18:09 EST Warning   Switch status changed from HEALTHY to DOWN.  Switch    71   1    FW-1424   CCFICON1 

Wed Nov 04 2009 12:18:09 EST Warning   Switch status change contributing factor Power supply: 1 bad.  Switch    72   1    FW-1427   CCFICON1 

Wed Nov 04 2009 12:18:05 EST Error     PS 1 set to faulty, rc=2000e.     Chassis   70   1    EM-1034   CCFICON1 

Wed Nov 04 2009 11:43:39 EST Warning   Trace dump available (Slot 4)! (reason: FFDC) Chassis   61   1    TRCE-1001 CCFICON1 

Wed Nov 04 2009 11:43:39 EST Warning   Trace dump (Slot 4) was not transferred because trace auto-FTP disabled.    Chassis   62   1    TRCE-1004 CCFICON1 

Wed Nov 04 2009 11:43:39 EST Warning   Trace dump available (Slot 4)! (reason: FFDC) Chassis   63   1    TRCE-1001 CCFICON1 

Wed Nov 04 2009 11:43:39 EST Warning   Trace dump (Slot 4) was not transferred because trace auto-FTP disabled.    Chassis   64   1    TRCE-1004 CCFICON1 

Wed Nov 04 2009 11:42:35 EST Error     CP in Slot 4 set to faulty because CP ERROR asserted. Chassis   48   1    EM-1033   CCFICON1 



External Moderator
Posts: 4,985
Registered: ‎02-23-2004

Re: I need ideas why FICON fails to configure online on DCX-4S

Peter,

sounds worse than it is.

FW-1427

Message

<timestamp>, , <sequence-number>,, WARNING, <system-name>, Switch status

change contributing factor Power supply:



<Number Bad> bad.

Probable Cause Indicates that the switch status is not in a healthy state. This occurred because the number of

faulty power supplies is greater than or equal to the policy set by the switchStatusPolicySet

command.

Recommended Action

Replace the faulty power supply.

Severity WARNING

Is a Power Supply bad ? or any light is amber ? If no, i assume the value in switchstatus are wrong, you can set it whit the command "switchstatuspolicyset"

FW-1433

Message

<timestamp>, , <sequence-number>,, WARNING, <system-name>, Switch status



Probable Cause Indicates that the switch status is not in a healthy state. This occurred because the number of

faulty CPs is greater than or equal to the policy set by the switchStatusPolicySet command. The CPs

are non-redundant.

If you power cycle a Brocade 24000 chassis in dual-domain configuration, and then reset the

micro-switch of the active CP before the heartbeat is up, this will cause both CPs to come up in a

non-redundant state.

Recommended

Action

Run the firmwareShow command to verify that both CPs have compatible firmware levels. Run the

firmwareDownload command to install the same level of firmware to both CPs. Replace any faulty

CPs.

If you reset the micro-switch (the latch on the CP blade) on the active CP before the heartbeat was

up on a power cycle, and the CPs came up non-redundant, then you should reboot the CPs again to

clear the problem.

Severity WARNING

check whit "firmwareshow" if both cp have a correct FabricOS loaded and active

FW-1439

Message

<timestamp>, , <sequence-number>,, WARNING, <system-name>, Switch status

change contributing factor Switch offline.



Probable Cause Indicates that the switch status is not in a healthy state. This occurred because the switch is

offline.

Recommended Action

Run the switchEnable command.

TRCE-1001

Message

<timestamp>, , <sequence-number>,, WARNING, <system-name>, Trace dump

available< optional slot indicating on which slot the dump occurs >! (reason:

<Text explanation of what triggered the dump. (PANIC DUMP, WATCHDOG EXPIRED,

MANUAL, TRIGGER)>)



Probable Cause Indicates that trace dump files have been generated on the switch or the indicated slot. The reason

field indicates the cause for generating the dump as one of the following:

PANICDUMP generated by panic dump

WATCHDOG EXPIRED generated by hardware watchdog expiration

MANUAL generated by the tracedump -n command

Recommended

Action

Run supportFtp (as needed) to set up automatic FTP transfers; then run the supportSave command

and contact your switch service provider.

Severity WARNING

Try "supportftp" to move all Dump file to FTP server, whit option -R will all dump deleted from CP.















TechHelp24

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.

Click to Register
Download FREE NVMe eBook