Contents: SQL Tools, Administration
Anchor | ||||
---|---|---|---|---|
|
SQL Tools
...
Code Block |
---|
[tersk02]:u/cd/greg> source /afs/slac/g/cd/soft/dev/script/ENVS.csh [tersk02]:u/cd/greg> source !:1:h/oracle_env.csh /afs/slac/g/cd/soft/dev/script/aidaSetEnv.csh [DEV|PROD|LCLSDEV|LCLS] |
Make sure you have TWO_TASK set appropriately:
[tersk02]:u/cd/greg> printenv TWO_TASK
DEV=AIDADEV@SLACPROD
PROD=AIDAPROD@SLACPROD
LCLSDEV=AIDA@MCCQA
LCLS=AIDA@MCCO
(See these defined in /afs/slac/g/cd/soft/ref/package/aida/common/script/aidaSetEnv.csh)
WHEN FIRST ADDING A NAME, USE "DEV". That will put the new name in AIDADEV. So you will then check that the AIDADEV network can get a value for that name. After you've verified it all works in AIDADEV, repeat for AIDAPROD, test again, and finally put the name in MCCQA. After the Synch to MCCO, AIDALCLS will be able to perform the query for that name.
Then enter a sql interpreter like sqlplus. You give the Aida database on which you will be performing operations. On tersk this is:
Code Block |
---|
sqlplus AIDADEV/<password> or sqlplus AIDAPROD/<password> // same password for AIDADEV and AIDAPROD |
You can also use emacs' sql-oracle mode. You'll need to ask an Oracle administrator for the password.
Adding a Name
First, find out the AIDA Server id number for which you'll be adding the name. To do that eithe rjust look in AIDA_SERVICES table yourself, or use the script show_services.
Second, add a row to AIDA_NAMES and AIDA_DIRECTORY to implement the new entity, per the schema defined in the Aida Schema. Effectively, do what is defined in /afs/slac/g/cd/soft/ref/package/aida/common/script/add_IA.sql or add_IAT.sql.
Management SQL scripts
This subsection describes the use of some of the many SQL scripts that you can find in /afs/slac/g/cd/soft/package/aida/common/script/.
...
To show names and ID numbers of all Aida services in the Aida directory service:
Code Block |
---|
@show_services |
This can be executed from the command line with "show_services".
To Add a Name to Aida.
To add a name to Aida, run add_IA.sql, giving the number of the Aida service to which you want to add the name, plus the instance and attribute of the name:
...
You can find all the existing Aida service names with show_services.sql. Add the name first to AIDADEV, then test; if successful, add the name to AIDAPROD.
add_IA can be used from the command line to execute add_IA.sql.
To Add a Name with a Transform to Aida
...
Note the use of quotes - you must quote the transform so it's a single argument to add_IAT.sql, and if teh the transform itself includes quotes as this one does, then they must be double-quoted in sqlplus.
Add the name first to AIDADEV, then test; if successful, add the name to AIDAPROD.
Add a transform that includes a replacement pattern
The AIDA Name Server supports pattern replacement, to allow a transform to be written using the literal strings &instance and &attribute to stand for the instance and attribute of the aida name of the transform. This allows the same transform to be used for many instances and attributes. Note that, for efficiency of runtime processing, transforms that require pattern replacement must have '/' as their first character, so the AIDA directory service knows only to do the expensive pattern replacement after examining the first character. The second thing to note when adding transforms to the AIDA directory service with SQLPLus (as add_IA does), occurrances of & must be escaped. Putting these two considerations together here's an example: the following name and transform will allow AIDA to acquire the SUML_M (the sum of the effective lengths up to the device) associated with XCOR:IN20:121. And the same transform would work for all devices.
Code Block |
---|
add_IA 5 XCOR:IN20:121 suml_m "/select suml_m from
lcls_infrastructure.v_lcls_elements_report where epics_device_name = ''\&instance''"
|
This will result in the following query text sent to service 5 (the Oracle data provider):
Code Block |
---|
select suml_m from lcls_infrastructure.v_lcls_elements_report where epics_device_name = 'XCOR:IN20:121'
|
Verify presence of a name
...
Code Block |
---|
@show_IA 'BPMS:IA20:235' 'twiss' |
show_IA can be executed from the command line. To use wildcards, or get the transform etc, use the APEX interface.
Show everything about an AIDA name
Use show_all. This example is more sophisticated, showing how to get information about all the twiss names of all the instances in a file, and recording results to a file:
Code Block |
---|
cat ~/checkDevices27Oct2008.txt_Aida_NO | xargs -I {} -t show_all {} twiss >& show_output.txt
or, using awk to select a field from a file
awk '{print $1}' < checkDevices27Oct2008.txt_lexid_DPonDEV | xargs -I {} -t show_all {} twiss
|
To Remove a Name from Aida.
...
Remove the name first from AIDADEV, then test, then it from AIDAPROD too.
remove_IA can be executed from the command line.
To Change the Transform of a Name
...
Change the transform first in AIDADEV, then test; if successful, change the transform in AIDAPROD.
replace_IA can be used at the command line.
To Add a Transform to an Existing Name
To make Aida send a transform to the data provider for a given instance-attribute name, use update_IAT.sql, just as you would have had the name had a transform associated with it before.
Add the transform first in AIDADEV, then test; if successful, add the transform in AIDAPROD.
Anchor | ||||
---|---|---|---|---|
|
Execute aida_xalServ_names_update by hand.
aida_xalServ_names_update
looks at the XAL model devices in MACHINE_MODEL schema on the instance on which it's executed, and updates the given AIDA schema on that instance, with any changes it finds.
So, it's run automatically by ModelManager after a model has been uploaded that contains new devices or devices that have changed name.
If it has to be run by hand
Log into MACHINE_MODEL on the instance whose AIDA schema you want to update, eg AIDADEV, AIDAPROD, MCCQA:
Code Block |
---|
SQL> SET SERVEROUTPUT ON
SQL> set timing on
SQL> SPOOL RUN_AIDA_XALSERV_NAMES_UPDATE_MCCO.LOG
|
Execute it, with last arg 'P' for pending.
Code Block |
---|
SQL> EXEC ONLINE_MODEL_PKG.AIDA_XALSERV_NAMES_UPDATE('AIDA', '202', '202', 'P');
SQL> SPOOL OFF;
|
Then check the results. If the AIDA schema changes check out, then commit the change.
Code Block |
---|
SQL> select n.instance, n.attribute, s.name, n.transform from
aida.aida_names n, aida.aida_directory d, aida.aida_services s where
n.instance like '&1' and n.attribute like '&2' and n.id = d.name_id and d.service_id = s.id;
SQL> commit;
|
Check the results actually by getting model.
Execute aida_symbols_names_update by hand.
aida_symbols_names_update
updates attributes like element.effective_length and element.S, used by Henrik's model applications.
At the time of writing, it is still only executed by hand following a presumed update in LCLS_INFRASTRUCTURE. That has to be fixed.
If it has to be run by hand
Log into MACHINE_MODEL on the instance whose AIDA schema you want to update, eg AIDADEV, AIDAPROD, MCCQA:
Code Block |
---|
SET SERVEROUTPUT ON
set timing on
SPOOL RUN_AIDA_SYMBOLS_NAMES_UPDATE_MCCO.LOG
|
Execute it, with last arg 'P' for pending. Replace the schema name with the appropriate one for the instance on which you're running.
Code Block |
---|
EXEC ONLINE_MODEL_PKG.AIDA_SYMBOLS_NAMES_UPDATE('AIDA', '5', '5', 'P');
SQL> SPOOL OFF;
|
Then check the results. If the AIDA schema changes check out, then commit the change. Eg, on MCCO the aida schema is called AIDA, so:
Code Block |
---|
define I='YCOR:LI30:900'
define A='element.effective_length'
select n.instance, n.attribute, s.name, n.transform from
AIDA.aida_names n, AIDA.aida_directory d, AIDA.aida_services s where
n.instance like '&I' and n.attribute like '&A' and n.id = d.name_id and d.service_id = s.id;
SQL> commit;
|
Check the results by aidaget. Eg: YCOR:LI30:900//element.effective_length
. If you have a file of the names for which to get data, you can check them all with something like;
Code Block |
---|
cat uniq_mag_names.txt | xargs -t -I{} aidaget {}//element.S
|
Synchronize MCCQA to MCCO
On a Solaris host, after AIDA development environment setup, execute:
Code Block |
---|
aida_validate_and_sync_instances.sh 'what, eg SLC Names'
|
or, in SQL:
Code Block |
---|
SQL> @aida_qatoo_sync.sql 'what, eg SLC names'
|
Administration
This section presents documentation useful to the Database administrators and developers responsible for the maintenance of the Aida directory service. This information is at a lower level than the SQL Tools presented above.
The Aida Directory Service database schema AIDA_Dir_Service_Db_Schema.pdf.
The Schema DDL of AIDADEV (identical to AIDAPROD) contains a listing of all table definitions, triggers and Stored Procedure listsings listings used by AIDA.