CA Service Management

Hi folks,

Is someone have an idea on how to mimic the ticket reference number allocation on CI?

I want to automatically add a Sequence number at the end of a CI name.

Any idea ?

Thanks.

Hi,

I think that you should avoid same named CIs by another method but...

I tried to get all of my CIs by SPEL, query that return 26727 CIs was acquired for 3 seconds.

05/26 11:25:26.40  sqlagt:select8       5976 SIGNIFICANT  prov_base.c           1010 SQL Query has returned (5000) rows for (domsrvr/superuser). Clause (SELECT ca_owned_resource.last_update_date, ca_owned_resource.own_resource_uuid FROM ca_owned_resource WHERE ca_owned_resource.own_resource_uuid IS NOT NULL) Input (<None>)
05/26 11:25:27.27  sqlagt:select8       5976 SIGNIFICANT  prov_base.c           1010 SQL Query has returned (10000) rows for (domsrvr/superuser). Clause (SELECT ca_owned_resource.last_update_date, ca_owned_resource.own_resource_uuid FROM ca_owned_resource WHERE ca_owned_resource.own_resource_uuid IS NOT NULL) Input (<None>)
05/26 11:25:28.18  sqlagt:select8       5976 SIGNIFICANT  prov_base.c           1010 SQL Query has returned (15000) rows for (domsrvr/superuser). Clause (SELECT ca_owned_resource.last_update_date, ca_owned_resource.own_resource_uuid FROM ca_owned_resource WHERE ca_owned_resource.own_resource_uuid IS NOT NULL) Input (<None>)
05/26 11:25:28.93  sqlagt:select8       5976 SIGNIFICANT  prov_base.c           1010 SQL Query has returned (20000) rows for (domsrvr/superuser). Clause (SELECT ca_owned_resource.last_update_date, ca_owned_resource.own_resource_uuid FROM ca_owned_resource WHERE ca_owned_resource.own_resource_uuid IS NOT NULL) Input (<None>)
05/26 11:25:29.66  sqlagt:select8       5976 SIGNIFICANT  prov_base.c           1010 SQL Query has returned (25000) rows for (domsrvr/superuser). Clause (SELECT ca_owned_resource.last_update_date, ca_owned_resource.own_resource_uuid FROM ca_owned_resource WHERE ca_owned_resource.own_resource_uuid IS NOT NULL) Input (<None>)
05/26 11:25:29.93  sqlagt:select8       5976 SIGNIFICANT  prov_base.c           1025 SQL Query has finished (26727) rows for (domsrvr/superuser).  Clause (SELECT ca_owned_resource.last_update_date, ca_owned_resource.own_resource_uuid FROM ca_owned_resource WHERE ca_owned_resource.own_resource_uuid IS NOT NULL) Input (<None>)
05/26 11:25:29.94  spelsrvr             5200 SIGNIFICANT  TEST                  3 Amount = 26727

SPEL code was:

send_wait(0, top_object(), "call_attr", "nr", "sync_fetch", "STATIC", "id is not NULL", -1, 0);
logf(SIGNIFICANT,"Amount = %d",msg[1]);

So this method can cause system lag.

Another way is to create new table, where will be CI reference and row id (which will be a sequence postfix).

But here is a lot complexities, how to prevent manual sequence edit and to avoid sequence duplication.

My first idea was to create a new table with a ID stored in it. So you tell me that i should avoid this plan, but right now we're using a Excel file to keep sequence in it, the technician have to check in the file to get the correct sequence when creating a new CI. Painfull.

So, any advice? Or i'm going with the new table approach?

And, if i look into the ca_owned_resource, there's a field named own_resource_id, if this field was accessible via SpelCode it'll be as easy as 123. I think i will create a SQL trigger to concat this value to the resource_name.

So if have only few CIs and you keep them actual manually, you can create new attribute or use OOTB field like asset_num.

Then define unique asset nums for all of your CIs.

Your SPEL basic variant could look like:

    mod file:

MODIFY nr POST_VALIDATE z_concat_name() 1337;

     spl file:

nr::z_concat_name(...)
{
  if( (!is_null(name)) && (!is_null(asset_num)) ){
    int result;
    result = sindex(name, asset_num);
    if (result==-1) {
      send_wait(0, this, "call_attr", "name", "set_val", format("%s.%s",name,asset_num), "SURE_SET");
    }
  }
}

I also would hide sequence postfix on detail edit but this is optionaly.

Pretty cool but even if i use this method, there's no mechanism to auto increment my sequence number. And i can't use the first spelCode you provided for the same reason you said, performance issue.

SDM is already generating an unique integer ID in own_resource_id. But i can't find this field in the dataModel, but it's in the database (table ca_owned_resource).

And in my case, some CI Class would require an generated suffix ID, but some not. So i'm going for : Create a CI, concat the ca_owned_resource.own_resource_id value to the resource_name via SQL Trigger, only if the boolean value in the CI's class allow me to do so, Disable the resource_name field, based on a checkbox in the CI's class. No sequence edit, no duplicate.

Perhaps this is not the best way to achieve it, but i'm running out of time and idea :/

In our environment we haven't any own_resource_id (only own_resource_uuid which is UUID (binary format)).

Okay, going back to the new table:

For example call it z_nr_concat;

Add SREL to nr called "nr";

Here is another script from my script tray :)

This will fill z_nr_concat table with CIs, so you can use id from this table.

nr::z_nr_concat(...)
{
  object zfound, zobj;
  int zcount, result_id;


  if(is_null(name)){
  logf(SIGNIFICANT, "Can't proceed. Values are empty.");
  return;
  }


  send_wait(0, top_object(), "call_attr", "z_nr_concat", "sync_fetch", "STATIC", format("nr=U'%s'", id), -1, 0);
  zcount = msg[1];

  if (zcount == 0) {
  object group_leader, new_nr;
  send_wait(0, top_object(), "get_co_group");
  if (msg_error()) {
  logf(ERROR, "New NR: error '%s' on '%s'", msg[0], peristent_id);
  return;
  }
  group_leader = msg[0];
  send_wait( 0, top_object(), "call_attr", "z_gts", "get_new_dob", NULL, NULL, group_leader);
  if (msg_error()) {
  logf(ERROR, "New NR: error '%s' on '%s'", msg[0], peristent_id);
  return;
  }


  new_nr = msg[0];
  new_nr.nr = id;
  send_wait(0, group_leader, "checkin");
  if (msg_error()) {
  logf(ERROR, "New NR: error '%s' on '%s'", msg[0], peristent_id);
  send_wait(0, group_leader, "uncheck");
  return;
  }
  result_id = new_nr.id;
  } else {
  zfound = msg[0];
  send_wait(0, zfound, "dob_by_index", "DEFAULT", (int)(zcount-1), (int)(zcount-1));
  zobj = msg[0];
  result_id = new_nr.id;
  }

  int result;
  result = sindex(name, format(".%d",result_id));
  if (result==-1) {
  send_wait(0, this, "call_attr", "name", "set_val", format("%s.%s",name,result_id), "SURE_SET");
  }
}

This is not tested but if it works you'll get autoincrement that will attach to your name on CI modification.

Works fine, needed some correction. Thanks a lot.

Here is the working version if someone want it :

I can't figure out how to embbed code, sorry

object zfound, zobj;

int zcount, result_id;

if(is_null(name))

{

logf(SIGNIFICANT, "Can't proceed. Values are empty.");

return;

}

send_wait(0, top_object(), "call_attr", "z_nr_concat", "sync_fetch", "STATIC", format("nr=U'%s'", id), -1, 0);

zcount = msg[1];

if (zcount == 0)

{

object new_nr;

send_wait(0, top_object(), "get_co_group");

if (msg_error())

{

logf(ERROR, "New NR: error '%s' on '%s'", msg[0], peristent_id);

return;

}

group_leader = msg[0];

send_wait( 0, top_object(), "call_attr", "z_nr_concat", "get_new_dob", NULL, NULL, group_leader);

if (msg_error())

{

logf(ERROR, "New NR: error '%s' on '%s'", msg[0], persistent_id);

return;

}

new_nr = msg[0];

new_nr.nr = id;

send_wait(0, group_leader, "checkin");

if (msg_error())

{

logf(ERROR, "New NR: error '%s' on '%s'", msg[0], peristent_id);

send_wait(0, group_leader, "uncheck");

return;

}

result_id = new_nr.id;

}

else

{

zfound = msg[0];

send_wait(0, zfound, "dob_by_index", "DEFAULT", (int)(zcount-1), (int)(zcount-1));

zobj = msg[0];

result_id = zobj.id;

}

int result;

result = sindex(name, format("-%d",result_id));

if (result==-1)

{

name=format("%s-%d",name,result_id);

}

sounds great :)

you need to switch to advanced editor to embbed code,

also I reccomend to use SET_VAL except of simple "=".

send_wait(0, this, "call_attr", "name", "set_val", format("%s-%s",name,result_id), "SURE_SET");

Modified. Thanks.

Advance mode? Where?

have searched for this function for a long time ))

ps:sorry for interface translation it cann't be switched

Oh Okay you can't access this advanced editor when answering from the home page :)
Thank you!