Why SCTP in Linux Kernel ?

Well, everyone is familiar with TCP and UDP, why still need SCTP?

SCTP provided new features as well as existed features in TCP.
Core Features of SCTP:
[1] Message Oriented
[2] Multihoming
[3] Streams

Besides that SCTP have features same as TCP:
[1] Congestion Control
[2] Selective ACK
[3] Message Fragmentation
[4] Bundling

When we need SCTP?

Let's consider a "Network Application" which operates by exchanging simultaneously, short, similar sequences of data continuously, such as client-server database application.
What will be the traffic condition for this "Network Application"?

Resource from: http://old.lwn.net/2001/features/OLS/pdf/pdf/sctp.pdf

SETUP CMM debug flags for 4.x, 5.x, and 6.x firmware

# Edit /etc/debug.cfg file to include data below to turn on debug flags:
FRU_DEBUG = 1
EKEY_DEBUG = 1
REA_DEBUG = 1
IPMB_DEBUG = 1
FAILOVER_DEBUG = 1
SYNC_DEBUG = 1 

Julesburg software pre and post script execution

Julesburg software pre and post script execution:

By default, Julesburg software is preset with /etc/cmm/scripts/pre_cmm_startup script to execute commands for CMM initial start up.

Besides the default script, there are 3 more script that can be preloaded during WW/CMM operation to handle automated scripting or commands execution.

Default:

/etc/cmm/scripts/pre_cmm_startup

Others:

/etc/cmm/scripts/post_cmm_startup
/etc/cmm/scripts/pre_cmm_shutdown
/etc/cmm/scripts/post_cmm_shutdown

Each of the scripts above executed at different time of CMM operations.

pre_cmm_startup: Executed once WindRiver Linux has boot up but before Julesburg is brought up.

post_cmm_startup: Executed when Julesburg/CMM just boot up or initialized and running.

pre_cmm_shutdown: Executed when Julesburg/CMM received terminate CMM/Julesburg signal but CMM not yet stop.

post_cmm_shutdown: Executed when Julesburg/CMM received terminate CMM/Julesburg signal and after CMM stop.

These scripts are useful when customers or engineers want to execute scripts/commands/creating logs and etc at the 4 time phase explained above.

UCM ClearCase View creation and removal (Linux/UNIX view)

This document: ClearCaseUCM_howto_createview.docx

One time step for Linux like utilities in Windows 7

1.       Unzip “custom-path.zip” to C:\

2.       Goto My Computer > Properties

3.       Select “Advanced system settings”

4.       Select “Environment Variables”

5.       Edit user variable “Path”, include the “C:\custom-path;”

6.       Click ok, ok.

7.       To take user environment variable effective, just logout and log in again.

STEPS to create UCM ClearCase view and remove UCM ClearCase view

1.       Identify what Project and Stream you going to work on using “ClearCase Project Explorer”.

2.       Goto C:\custom-path-radisys and click “Create_UCM_Linux_view.bat” (you create this as shortcut to your desktop or any other place)

3.       After created the view in Linux/UNIX, the screen below pop up to synchronize it to Windows ClearCase. (to allow access/work on Linux/UNIX view from Windows)

4.       Select the view has just created and click “Import”.

5.       Warning below pop up, just click “ok”

6.       After that window below pop up, and edit \\mycc01 to \\mynas01, then click “ok”.

7.       After finish import, click “Close” and view is created in Linux and visible in both Windows and Linux now.

8.       To confirm UCM view created to correct stream, goto “ClearCase Project Explorer” and select the stream’s properties.

9.       All view attached to the stream will be listed in “view” tab.

HOW to remove UCM view:

1.       Login 10.111.2.38 or mycc01 server

2.       List your view name, “cleartool lsview | grep ”

3.       To remove, “cleartool rmview –tag ” and it is done.

Upgrade old CDM to new CDM using latest WW/JB firmware

This guide is to help on how to flash CDM(before 8.1.1.01) when using WW/JB after 8.1.1.01.

Issue/problem:

After WW boot up in chassis (6014/NEC), it not able goto Act/Stdby and at the same time rsys-ipmitool not able to flash/read CDM (error message device ID not found)

Solution:

1. Stop CMM (command: "cmm stop")

2. Manually input / tell IPMC what chassis are you using now with command below:

   i) NEC (CMM will reboot after execute command below)

      rsys-ipmitool raw 0x2e 0xc0 0xf1 0x10 0 5

   ii) 6014/6016

      rsys-ipmitool raw 0x2e 0xc0 0xf1 0x10 0 6

3. Once WW rebooted, login again and stop the cmm.

4. In CMM1, Verify you can read the CDM now by executing command

    rsys-ipmitool -t 0x20 -m 0x10 fru read 1 FILENAME.bin

5. If can read, then proceed with flashing new CDM in CMM1

    rsys-ipmitool -t 0x20 -m 0x10 fru write 1 NEWCDM.bin

6. Reboot/power cycle whole chassis and wait the WW/JB boot up and go into Act/Stdby state.

Timnath: downgrade 7.x (JB) to 5.x (Timnath) on Colorado Springs (CMM hardware)

Downgrading from 7.x JulesBurg to 5.x (Timnath) CAN NOT use "cmmset -d update" CLI command.

The steps below need to be followed in order to downgrade:

1. copy \\10.111.2.37\root\tftpboot\ColoradoSprings\downgradeFrom7.x to home folder and rename to downgrade. (there is character limitation on cmmset -d update)

2. Transfer (SCP/FTP) into CMM /usr/local/data/temp

3. Downgrade the CMM with command below:

cmmset -d update -v "/usr/local/data/temp/___CMM_5.2.8.75_DOWNGRADE__
_ force"

Configure NIS server

1. Login as root

2. in terminal, type "authconfig"

3. select NIS and click [NEXT]

4. Domain=cybertech

5. Server=cybertech01.cybertech.com

6. ypwhich to verify machine is bound to cybertech01

7. ypcat passwd | more to verify machine is bound and login accounts.

Just run "ypwhich" command after authconfig to verify that you are bound to "cybertech01".
I also do a "ypcat passwd| more" this will verify that you are bound AND youcan see login accounts.