SIAM Code Daily Analysis and Testing

by:

Karen A. Salamy

(prepared 5/27/04)

This document describes how to generate and install SIAM software releases, and suggests standard methodology for SIAM output analysis.

Preparation:

1)@3:00PM ASK everyone to check-in their new code

2)In /home/salamy/siam perform a:#cvs update -d

(can use GCVS GUI to do this)

3)Run a:# make cleanin /home/salamy/siam

4)Run:# makein /home/salamy/siam

5)Hope everything builds properly, otherwise talk to user whose code is causing problems.

6)Perform a TAG in CVS with the current date (e.g., apr30)

First run list of last tag number (from ~/siam):# cvs log Makefile | more

Then Tag with new number for day- cvs rtag tag_month_day_year siam(from any directory is okay)

Example:# cvs rtag apr30_2004-2 siam

7)Export the new TAG to a /home/salamy/tmp directory

# cd /home/salamy/tmp(create the /tmp directory if it does not exist)

# rm –rf *

# cvs export –r apr30_2004 siam(use the tag of the day – e.g., apr30)

8)Manually set parameters (BE CAREFUL!Close window when done as this will affect your environment settings)

 

# export JAVA_DEV_ROOT=/home/salamy/tmp/siam

# export SIAM_HOME=/home/salamy/tmp/siam

9)cd into siam/ and make the following directories:

# cd siam

# mkdir classes

# mkdir ports

10)Now, ready to perform a make:

# make

# make cimtpucks

# make distribution

11)# cd ../tmp(one level above siam)- ASK everyone if it is okay to blow away old siam/.

Now have a deployed.tar.gz file for transferring first to siam-portal-2 then to mooring

From greenflash:/home/salamy/tmp, copy all *.tar.gz files over to the portal:

# scp *.gz ops@siam-portal-2:/home/ops

Next, log in to siam-portal-2

# ssh ops@siam-portal-2(log in to siam-portal-2 for transfer to mooring)

Next, transfer compressed *.tars to mooring:

# cat /etc/hosts(Gives list of hosts – verify mooring is in the list)

# ping mooring (make sure the node is alive)

Follow MMC node update procedures as described  here .

Now ready to go but NEED TO SET UP PORTAL – Usually siam-portal-2

Copy shore.tar.gz file to /home/ops on siam-portal-2

From salamy@greenflash:/home/salamy/tmp:

# ssh ops@siam-portal-2(login)

# rm –rf siam

# tar xvzf shore.tar.gz

# cd siam

# mkdir logs

# cd logs

Next, Verify no other portal running on portal (siam-portal-2) system:

# ps –eaf | grep jdk

If one or more exists, kill first jdk process running with:kill

Start portal:

Publish in background:# portal –publish >&portal.log &

OR in foreground:# portal –publish mooring-cimt | tee portal.log

Next, CHECK TO SEE IF YOU NEED TO BURN PUCKS!!!! 

Before starting the test, create a list of the changes to the evening’s code since it was last tagged to send out to the rest of the SIAM team in e-mail.

Skip this step if no code updates were performed for the day.

From /home/salamy/siam type:

# siamChanges 'mm/dd/yyyy hh:mm:ss' > junk

Where an example of the date/time of last tag is:‘05/07/2004 16:00:00’

Piping out data into “junk” file makes it easier to read changes.E-mail these to the SIAM Users / SCRUM list.

Start evening test: Usually from either greenflash or etna as root@mooring

# ssh root@mooring

# reboot (or halt)

Dial modem – 831-769-1050 to bring system back up.

Watch /logs directory on portalhost (siam-portal-2) (in first window) to see if data is being logged.

NOTE:Most of the time, you will not manually enable the node.* file.This will be done by enabling the autostart parameter and rebooting.

To view the bring up of the instruments, login ops@etna and do the following:

# gosiam

# listPorts localhost –stats

STANDARD DATA RETRIEVAL:

To pull data off of system while it is running and is in its housing

From siam-portal-2:/home/ops

# lease 1800 mooring(obtains 30 minute lease)

(This length of time is good for weekend data retrievals if the node was set to INFO mode for test.1200 should suffice for weekday retrievals

in INFO mode. Double the lease time if DEBUG mode was invoked as it produces more comments and larger files that take longer to download

to the portal.)

Wait for prompt to return, then:

# ssh ops@mooring(NOTE: If tests are not running, simply start here for data retrieval)

# cd /mnt/hda/siam/logs

# tar czf nodedata.tar *(NOTE: If the node is running, first copy the node log (node.*) to another name as you cannot tar a file

that is being actively written to.You will need to exclude the active log file in the tar command in this case.) 

# exit

From siam-portal-2:/home/ops

# scp ops@mooring:/mnt/hda/siam/logs/nodedata.tar .

To transfer data over to morgana:

# ssh ops@morgana

# cd archive

# mkdir Tag_test_date-overnight(Example:mkdir may06_2004-overnight)

# cd Tag_test_date-overnight

# scp ops@siam-portal-2:/home/ops/nodedata.tar .

Also, move the portal.log document into ops@morgana:/home/ops/archive/tagged_run_of_day:

# scp ops@siam-portal-2:/home/ops/siam/logs/portal.log .

///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

DO ONLY WHEN SYSTEM IS NOT PACKED IN ITS HOUSING (Testing in lab with Ethernet connection):To stop a test at any time do the following using a console port.Plug DB9 cable into Etna’s serial port. From etna type:

# minicom

# lease

# exitNode localhost

To bring up the Ethernet on node following a shutdown:

# ps –eaf | grep j9

# poff

# ifdown eth0 (just to be sure)

# ifup eth0

To bring up Ethernet on system that is still running do the following at the mooring-cimt prompt on Etna:

# cpuAwake

# ifup eth0

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

Daily analysis of tests:

The SIAM node puts out *.dat and *.idx binary files prefaced by ISI ID for each instrument.

Also puts out a node.* ASCII file that will be used to analyze test.These were placed into the archive directory on morgana in the Standard Data Retrieval step above:

# ssh ops@morgana

Change directory into the newly created directory (last step) of the most recent overnight tests: ops@morgana:/home/ops/archive/Tag_test_date-overnightand grep the node.* ASCII file for ERROR and WARN.

# grep ERROR node.* > errors

# grep WARN node.* > warns

Compile all ERROR and WARN statements into an e-mail daily to the SCRUM group (oreilly, bobh, headley, mrisi, graybeal, kgomes, and chma).