Re: [foxboro] ICC to CP disconnect

Matt,

ICC communicates with the CP, but uses the station workfiles for
information.    Look in your /opt/fox/ciocfg/CPLBUG directory where
CPLBUG is your CP's letterbug.  Look for CPLBUG.wf and COMP.O files
there - do they exist?

Also, generate a CSA save and check to see if your problem CP's
compounds and blocks are listed.  To do this, on your CSA server:
  cd /usr/fox/csa
  mkdir save_20Feb08
  CSA_Save ./save_20Feb08 (or CSA_Save /usr/fox/csa ./save_20Feb08 if
you get a "No destination" error from the first command)
  cd save_20Feb08
  more CPLBUG (where CPLBUG is the letterbug of your problem CP)
You should see a list of all the compounds and blocks in that CP.  Is
this information present?

If the CSA data and your station workfiles are all missing I'd suggest
that an Initialize was accidently done on this CP.  Check your system
monitor messages.

Foxboro's official policy has always to fix problems with CP-workfile
mismatches by doing an initialize/loadall of the affected CP.

Doing an upload will not help you - this will not build blocks, it
only uploads the values of the settable parameters in your workfiles
from the running CP.

Because your saveall gets it's data from the same place ICC does
(station workfiles) your savealls will be empty for this CP starting
from when the problem occurred.

For your lab system I'd recommend doing the initialize/loadall of the
CP.  If you are concerned about changes since the last successfull
saveall was done, save a copy of your checkpoint file
(/usr/fox/sp/files/DBCPLBUG.UC where CPLBUG is the CP letterbug) and
use dbvu to generate a block/parameter listing.  Compare that to an
iccprt listing of the CP's control database from the workfile after
the loadall and an upload.  Any differences can simply re-entered.  Or
you can send me the files and I can check for you - I have some
programs written (I may have to dig to locate them) to do block and
parameter level comparisons on these.  TAC may have something similar.

The unsupported method I'd probably try on a production system that
couldn't be rebooted would be to do that initialize/loadall on an
offline system.  Then copy the workfiles, .O files, and compound
subdirectories (with sequence and plb files) to the online AW and
check functionality.  If the saveall is very recent that should work
fine.  If it is older, you may have some mismatches between your CP
and your workfiles.  You can compare the data in the checkpoint file
(using dbvu) with the data in the workfile (using iccprt after doing
an upload).  If you were missing the CSA data also, you could merge
that data in from the CSA_Save done on your offline system after the
loadall.

Regards,

Kevin FitzGerrell

On 2/20/08, Gunter, Matt <Matt.Gunter@xxxxxxx> wrote:
> Here is the mystery:
>
> Using ICC I brought up one of our CPs to tweak a few Device parameters
> but found ... NOTHING ...  No blocks, no compounds, heck there weren't
> even any ECB blocks - only the empty primary.  This CP looks, for all
> intents and purposes, like it has been initialized and rebooted.
> However, using FoxSelect, I can see all the compounds and blocks on this
> same CP and the sequence blocks can run.
>
> Here is the question:
>
> How can I upload the database that exists in the CP to the ICC?  After
> searching the documentation on the electronic documentation CD, it looks
> like I have the option to 1) REALLY initialize the CP and perform a
> Load_All from a relatively recent Save_All, or 2) use the ICCupload()
> command in the ICCAPI.  However, if I take the second choice, will it
> rebuild the ECB blocks as well?  Documentation doesn't state.
>
> History:
>
> After doing a lot of work on our lab system (fortunately this is not on
> a system running product [Hmmm!!!]) I decided to perform a Delete &
> Undelete on all of the compounds in the different CPs that I had been
> working in.  There are eight compounds on the CP in question and I ran
> the D/U on only one of them.  However,  six hours later our system
> performed an automatic Save_All (cron job) and the Save_All files are
> basically empty.  The other oddity here is that seven of the eight
> subdirectories on the AW (Solaris 8/FV 9.0.1/I/A 7.1.1) containing
> sequence logic blocks for compounds are still in tact.  One is,
> inexplicably, missing (scary music here).
>
> Any thoughts on what might of happened (including the idiot driving the
> keyboard) and the best course of action, would be appreciated.
>
> Thanks
>
> Matt Gunter
> ATK Launch Systems
>
> P.S.  I have just discovered that one of the other compound
> subdirectories is missing.  This time, however, the compound is not only
> available to Fox Select, but also to the ICC.  It seems like the moral
> of the story is to NOT use Delete and Undelete.
>
>
> _______________________________________________________________________
> This mailing list is neither sponsored nor endorsed by Invensys Process
> Systems (formerly The Foxboro Company). Use the info you obtain here at
> your own risks. Read http://www.thecassandraproject.org/disclaimer.html
>
> foxboro mailing list:             http://www.freelists.org/list/foxboro
> to subscribe:         mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
> to unsubscribe:      mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave
>
>
 
 
_______________________________________________________________________
This mailing list is neither sponsored nor endorsed by Invensys Process
Systems (formerly The Foxboro Company). Use the info you obtain here at
your own risks. Read http://www.thecassandraproject.org/disclaimer.html
 
foxboro mailing list:             http://www.freelists.org/list/foxboro
to subscribe:         mailto:foxboro-request@xxxxxxxxxxxxx?subject=join
to unsubscribe:      mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave

Other related posts: