Messages

16

Documentation / Re: CMOD 10.1 ODWEK compatibility

« on: November 27, 2018, 01:13:19 AM »

Yah, we are trying to keep the ODWEK and SRV as close as possible, but what would you do in a case when you have a custom app which needs to serve two (in my case four) different instances of CMOD a kind of federation tool? However, the upgrade path of the servers is not possible to execute all at once (in one weekend). They are not very far of each other mostly just minor version, but in case to step up from 9.5 to 10.1 here you have major version change.

Anyway, I will post the solution if/when we get one ... 

Cheers,
 N.

17

Documentation / CMOD 10.1 ODWEK compatibility

« on: November 26, 2018, 07:02:21 AM »

Hi guys,

I just would like to share my findings in compatibility of ODWEK between CMOD Server 9.5 and 10.1.

https://www-01.ibm.com/support/docview.wss?uid=swg21392275

The ODWEK (the old "core") api used for search and retrieve of documents works as  expected. However,  using the ODWEK XMLparse method of ODServer, could give you a grief.

Since this is not a true Java API method, but rather masking with system call of arsxml utility, you will get incompatibility between server and client.
Basically the ODServer.XMLparse() method is not compatible in ANY of the combinations.

ODWEK 9.5 does not work with CMOD v10.1
nor
ODWEK 10.1 does not work with CMOD v9.5

Not sure if this is a bug or just an undocumented feature

Hope this helps to someone.
N.

18

z/OS Server / Re: Automatic User and Group synchronization with LDAP

« on: November 26, 2018, 06:43:33 AM »

Hi guys, thank you very much to both of you.
@Justin, yeh you are right I've realized that ... after  I have hit submit.

@rob, you are right I would need to find AD admin ( they all are "very" busy) and my knowledge of LDAP/AD is rather thin

Can I ask you for favor, in the example you have provided, would it be possible to outline the LDAP sub tree?

Thank you,
 N.

19

z/OS Server / Re: Automatic User and Group synchronization with LDAP

« on: November 22, 2018, 04:11:23 AM »

Hi guys,
sorry to hijack this thread, can some provide me with a sample of LDAP sub-tree which would demonstrate how to setup LDAP/AD to perform LDAP sync?
Creating CMOD Gorup(s)
Creating CMOD User(s)
Binding the respective user(s) to respective Group(s)

and the corresponding ARS_LDAP parameters mostly:
ARS_LDAP_USER_FILTER
ARS_LDAP_GROUP_FILTER
ARS_LDAP_GROUP_MAPPED_ATTRIBUTE


I would appreciate your help.
Thank you,
 N.

20

MP Server / Re: List of CLIP ID in TSM 7.1

« on: October 30, 2018, 03:10:52 AM »

It's not a perfect but you get an idea.

Replace:

• tmsinst1 password with yours
• ARCHIVE_NAME with your CMOD Instance name

Just a note:

• The ECS supports S3 protocol as well as CAS (Centera), you would be better of move to CMOD 10.1 and migrate from TSM/Centera to CMOD/S3.
• The TSM/Centera clips are *NOT* 1:1 with CMOD load files. The TSM stores some "preamble" in the file before it is stroed as clip in Centera.

Cheers,
 N.

Code: [Select]

#!/usr/bin/bash
if [ $# -ne 2 ] ; then
        echo "Usage: `basename $0` agid_name doc_name"
        exit 1
fi
AGID_NAME=$1
DOC_NAME=$2

db2 connect to tsmdb1 user tsminst1 using XXX_tsminst1Password_XXX >/dev/null 2>&1
trap 'db2 terminate' exit

SQL="select cast(assg.pvrpos as varchar(128) for sbcs data) from tsmdb1.AS_Segments assg join tsmdb1.AF_Segments afsg on
assg.volid=afsg.volid and afsg.bfid in (
select a.object_id from tsmdb1.archives a where a.filespace_name='/ARCHIVE_NAME/$AGID_NAME'
and a.LL_NAME = '$DOC_NAME') and assg.start=afsg.start"

echo $SQL | db2 -x +p
exit $?


21

MP Server / Re: migration strategy for this environment database db2 9.5 to oracle 12.1.0.2.0

« on: June 05, 2018, 08:13:33 AM »

To answer WHY to migrate from DB2 to Oracle, one good reason for all. Large companies like to streamline operations and that includes database services. To maintain a team of experts just for one DB2 while most of other DBs are Oracle and MS SQL based could be a nice way to reduce the cost.

I have done the migration of a smaller instance from DB2 Windows to Oracle Unix here are my key points:

• Upgrade CMOD first while keeping the DB2 to V10. This step should be easy to do
• Build a new vanilla instance of CMOD V10 using Oracle DB, make sure all is working well
• Upgrade CMOD first while keeping the DB2 to V10. This step should be easy to do
• Close all open tables using arstblsp -a1
• For migration of system tables (ARS%) I have wrote a perl DBI script to facilitate Export / Import, which would cover CLOBs (ARSAPP table!)
• As you have closed all seg tables, you can use arstblsp -a4 to create an empty schema table for each of your AGs at your Oracle CMOD. Note: In my case, I was not using SMS
• Duplicate the tables in Oracle database as required e.g.: create table segtable1 as select * from segtableN where 1=2;
• For segment tables use DB2 export to segtable.dat of DEL select * from segtable;
• Use sqlldr utility to load CSV exports from DB2 of your segtables. Note: you need to do a bit of fiddle with date and times to get it right.

Well that's all in a nut shell.

Disclaimer at the end  : use all above at your own risk, all notes I have typed are mine from my memories, so some minor tweaks could have been left out unintentionally.

Good luck and enjoy

22

MP Server / arssockd reload cached configuration

« on: June 05, 2018, 07:41:07 AM »

Guys, is there a way to reload (cached) configuration of arssockd on Unix without arssockd reboot? A kind of similar concept as to send a user signal to the arssockd process which will internally trigger configuration reload (such as storage set configuration for System Log etc.)?
Thx a lot,
 N.

23

MP Server / Re: adding and deleting the documents while doing migration

« on: March 15, 2018, 11:42:56 AM »

Hi,
not sure if I fully understand what you are trying to achieve, however as far as the deleted documents (post or during migration) IMHO you have two options:

• given you are using proper CMOD technique to delete or unload you can find relevant info by querying the System Log (or respective SLn table)
• The long route would be to do reconcile row counts of each of the migrated segment tables

The new arrivals; a great help would be to have all segments closed just before migration, so the "long running" migration can progress with static data not be "polluted" with new ingestion(s). Then after the "long migration" you can just re-run process on delta found in currently open segment.

Other route would be to subtract load_ids; from arsag.load_id on source system subtract the arsag.load_id on target system ; the difference gives you number of new ingestions to your source system since migration begun.
Sure, you can also get the information from System Log and/or System Load (if present).

Hope this helps or at least give you some pointers.
Cheers,
 N.

24

MP Server / Re: Cache Storage on new drive

« on: October 03, 2017, 08:26:24 AM »

Hi Gabriela,

I think you can try to explore windows command (almost forgotten) SUBST.

Associates a path with a drive letter.

SUBST [drive1: [drive2:]path]
SUBST drive1: /D

  drive1:        Specifies a virtual drive to which you want to assign a path.
  [drive2:]path  Specifies a physical drive and path you want to assign to
                 a virtual drive.
  /D             Deletes a substituted (virtual) drive.

Type SUBST with no parameters to display a list of current virtual drives.


I think it even works in user space, so no SA should be required, but don't take me granted 'm not Windoz guy

Cheers,
 N.

25

MP Server / CMOD MP Library server with zOS Object server

« on: October 03, 2017, 06:36:35 AM »

Guys,
I must admit haven't searched the forums in deep. 
So do blame me if this was already asked / answered elsewhere.

Anyone tried to connect CMOD Library server running on Linux with CMOD Object server running on z/OS mainframe?

I need to migrate huge CMOD (20+ years of data) from zOS to Linux farm. (If you would like to ask the obvious why: a. cost, b. the CMOD had never ran any expiry and OAM is configured to Never expire, so rather than expiring 60-70% of data, I would migrate what needs to be kept and can the rest)

So, the idea is:

• connect the two  Lib & Obj on MP to  Obj on zOS (if technically doable)
• script arsadmin retr/store from z/OS to MP, idea to run both ends from Linux

Thx for ideas,
N.

26

MP Server / Re: copying part of AGs to new server

« on: October 03, 2017, 06:16:10 AM »

Hi Michael, not sure if this is not a bit too late reply 
What you have describe in terms of copy / cloning the complete CMOD that works and yes I am using it just like you described.
However when are you trying to copy only partial database e.g. documents archive after the date X
I would scripted with DB2 export index table by index table (a dynamic SQL (sql generating sql) would help) and as part of export you can place restrictions (where clause) with your where mydate >= X.
then use arsadmin retrieve and store for matching doc_names.
One warning: Don't forget about resources and doc_name1 files they are not listed in index segments but rather in arsres. The doc_name1 are not listed at all in CMOD but are not vital, however nice to have  the doc_names (in case of AFP or PDFs)
Good luck,
N.

27

z/OS Server / Re: Running ARSMAINT via SSL to ARSSOCKD - Proof of concept

« on: September 25, 2017, 06:16:16 AM »

There could be a use case.

• If you have a Linux machine next to it, it is a bit easier (at least for me ) to script things on Linux then on z/OS via USS.
• It also could be a bit cheaper as it doesn't cost MIPS.
• Scripts could be re-used in MP environment as well

28

MP Server / ARSXML failed to create user with UID>49774

« on: August 01, 2017, 05:31:06 AM »

Guys, anyone noticed problem with creating (adding) a new user using arsxml when the last existing uid=49774 ?

Code: [Select]

<user password="XXXXXX" name="USER1234" description="A new user" fullName="USER1234" >
Failed with Err:

ARS7709I Adding user, USER1234
ARS7743E A user object named 'USER1234' already exists.
ARS7761I Add of user, USER1234 failed.


Forcing UID<49774 by finding a number range gap:

Code: [Select]

<user password="XXXXXX" uid="1001" name="USER1234" description="A new user" fullName="USER1234" >
Have worked well.

Any idea or known limitation?

29

Report Indexing / PDF Indexer with separate index data

« on: June 20, 2017, 12:46:13 PM »

Hi guys,
I have PDFs previously intended to be loaded by using Generic indexer (.IND) The PDF files do not contain PPDs.

However, I would like to use resource separation, in order to increase compression ratio, to do so I (guess) have to use the PDF Indexer.

Can you advise how to merge or instruct PDF Indexer to use the index data from .IND?

Note: I was looking into arspdoci tool, however 'm not sure if this is the right way (only way or no way) to go about.

Thx a lot

30

MP Server / Re: copying part of AGs to new server

« on: June 16, 2017, 05:58:47 AM »

You can also use arsdoc get and arsload to perform this.

I think that would be way too sloooow; assuming you have copied the database (respective segment tables), the arsadmin retrieve and store is the right way to go. As hint, (As Alessandro suggested once to me) use multiple doc_names on the command line (e.g. 1000+ should not be an issue on Unix) will increase the export speed.