03/24/2006

Start from the pages

Look at CESR run plan to have a general idea. From the <A HREF="http://w4.lns.cornell.edu/public/CESR/sched_2006e.gif"> 2006E </A>, we know that from Saturday March 4, 2006, CESR started deleavering the normal beam.

Open the CLEO E-Log page, we choose Calendar 2006. Click the March 4, looking at the column entitled "Sequence" with name

"Global DataTaking".

Digest this page

http://www.lns.cornell.edu/ccelog/search.pl?nocmt=1&amp;startDate=2006/3/4

Creat Time Table.py

python TimeTable.py -f ../dat/test/RunSum20060304.html

python TimeTable.py -f
"http://www.lns.cornell.edu/ccelog/search.pl?nocmt=1&startDate=2006/3/4"

03/26/2006

Go to CLEO E-Log page, search from Mar 4, 2006 to Mar 26, 2006. Save as RunSum20060304-20060326.html

python TimeTable.py -f ../dat/test/RunSum20060304-20060326.html

Lost some page. Consider day by day approach.

03/27/2006

Process 2006/3/4:

python TimeTable.py -f ../dat/test/RunSum20060304.html
215395  Comment: HV loss 80
215401  Comment: Good Short Run; Loss HV Telnet 81.
215404  Comment: Good Short Run; Loss comm for HV Telnet 82
215412  Comment: HV-box#81
215416  Comment: HV box#80
-----------------------------------------------
Run 215384 -- 215420 , Total time: 17:11:58
-----------------------------------------------
Content         ||Time  ||Percent
-----------------------------------------------
HV Box #80      |0:7:51 |0.761
HV Box #81      |0:7:9  |0.693
HV Box #82      |0:3:31 |0.341
-----------------------------------------------

Use the HV selected html file:

Go to CLEO E-Log page, at "Text Search", input "HV". search from Mar 4, 2006 to Mar 26, 2006.

Save as HV20060304-20060324.html

python TimeTable.py -f ../dat/html/HV20060304-20060324.html

03/29/2006

Continue on the script

python TimeTable.py -f ../dat/html/t.html

Found the comment block

03/30/2006

Work on the script

04/03/2006

Work on the script. Need more detailed example of evaluating the time.

Talk with Steve. Creat the three columns table first.

Done this.

python TimeTable.py -f ../dat/html/HV20060304-20060324.html  > ../dat/html/table1.txt

table1

04/04/2006

Categorize the table

Server crash, telnet lost, others

Creat this table list:

python TimeTable.py -f ../dat/html/HV20060304-20060324.html  > ../dat/html/table2.txt

------------------------------
Summary:
Server issue : 83 (% 48.5 )
Telnet issue : 11 (% 6.4 )
Other issue  : 77 (% 45.0 )
------------------------------

table2

04/07/2006

In Server issue:

HV losst connection with box#81
Should be categorized in the telnet?

In "Other Issues",

HV Server caught signal 11

What does this mean?

04/10/2006

Message from Ji Li:

Signal 11 means "Invalid memory reference" on a Linux machine. It is
usually caused by referencing invalid pointers.  As such, it should
definitely be categorized as "Server issue" (in rare cases, it can be
due to a defected hardware, for instance a bad RAM stick. there is
test we could do if that is what we suspect.). you can get the meaning
of all the 'signals' by using "man -S 7 signal".

From Basit:

We should sub categorize "server issues" to when during the run
we had problems. If the server crashes in the middle of the run
then it usually do not cause any time loss. If it crashes during
tranision or server fails to ramp up or down it causes time loss.

We should also quantify the the frequency of each HV box crashes.

From Dan:

Ok, things like HV Server caught signal 11 are generally not costing
us any running time, since the shifter can just restart the server
without any loss of HV.  It's certainly still useful to know how often
it happens, though, but that's something to keep in mind for
calculating how much luminosity we're losing. I guess the next step is
to look at which problems cost us luminosity and which
didn't. Sometimes server problems do cause a loss of data, so you have
to be careful in adding these things up.

Comment by Hubert during the Run 21660:

Looked up the switch port of c3lnx117 on netdisco: It is on a gigabit
port, while lnxon11 is on a 100MBit port. This might have an influence
on the better HVServer stability on lnxon11.

04/12/2006

python TimeTable.py -f ../dat/html/HV20060304-20060324.html  > ../dat/html/table3.txt

table3

Refer to the page of regex.