Messages

16

z/OS Server / Re: ARSYSPIN

« on: October 19, 2020, 08:47:15 AM »

They need their outputs in case there is an issue and they have a sweep at  5:30 and 19:30.
So, what happens if they have an issue with one of the outputs created at 5:29 or 19.29? The output will be available only in Mobius, right? If so, CMOD design should be changed to accommodate for this exception, right?

17

z/OS Server / Re: ARSYSPIN

« on: October 19, 2020, 05:43:15 AM »

ARSYSPIN is a started task.
This means, you can bring it down and keep it down with automation and restart it as well when you need to.

Typically, when we do maintenance, we bring down ARSYSPIN [and ARSLOAD]. And when maintenance is completed, we start the tasks again. All via automation.
You need to start talking with your mainframe automation team as well as z/OS Systems Programmers for how this can be achieved with the tools available in your shop.


I hope this helps,
Mehmet

18

z/OS Server / Re: z/OS CMOD Postprocessor documentation -- where is it?

« on: September 30, 2020, 10:12:44 AM »

Thank you Ed for the response.

We have been using the Linux approach as documented in the first link. There might be something else going on with our OMVS and z/OS environment. We
 will take another look to see if we can diagnose the problem. 

Update: Linux approach actually worked as expected. Maybe documentation should be updated one day to include z/OS in order to avoid confusion.

19

z/OS Server / z/OS CMOD Postprocessor documentation -- where is it?

« on: September 30, 2020, 06:06:16 AM »

We are trying to find some documentation for z/OS CMOD Postprocessor and we can't find it.

On the other side, there is documentation for Linux and Windows implementations here: https://www.ibm.com/support/pages/using-postprocessor-application-load-information-page-while-loading-data

...and for iOS implementation here: https://www.ibm.com/support/pages/using-postprocessor-parameters-content-manager-ondemand-i-i5os

Is there a similar technical reference for z/OS versions of CMOD? Maybe I am not able to remember where it is and I can't find it with internet search.

I appreciate it if anyone can help.

Bets Regards,
Mehmet

20

Windows Client / Re: Adding new accesses by multiple users

« on: September 29, 2020, 07:44:44 AM »

Have you considered using ARSXML ?
It could be a life saver in this instance when you have a lot of users to add.
It is very fast to execute and it could be a life saver when defining a lot of objects to CMOD such as users, groups, applications, etc... 

21

MP Server / Re: CMOD File Size Limit

« on: September 11, 2020, 11:13:12 AM »

PPD stands for PostScript Print Definition file.. I have not processed this type of a file so far in CMOD, and that's why, I looked it up from the internet. I am not clear what the data looks like.... AFP or XML like tagged text... or something else. A lot depends on the data and meta indexes for what you can and cannot do. 

Question: Is the file going to be stored and retrieved as a single document?

 If yes, you are probably out of luck and you will probably hit the 2 GB maximum document limit of CMOD.

 If no, assuming that the file contains smaller units of data (documents) delimited by some tags or meta indexes that will allow you to easily index them, and you may be in luck. As long as none of the individually delimited documents in the file exceeds 2 GB, you should be OK. I have processed files larger than  15 GB successfully and because the documents in the file were less than 2 GB, CMOD processing went fine.   

22

Report Indexing / Re: Changing indexed report to non index report

« on: September 01, 2020, 08:07:19 AM »

...
For full text search, all the work happens on the CMOD server.  And while this works great for modestly sized reports, if a user selects several multi-thousand page reports for full text search...  performance will suffer on the server, and the heavy I/O may impact other users.
...

This exactly has been the downside to the full text search.

However, let's say we have 10 large reports and 1 user. Full Text Search is not configured for the folder. User wants to search for a keyword in all 10 reports and the keyword is only available in 3 of them, but he doesn't know until all 10 reports have been opened/searched. What will the user do? He will bring all 10 reports to the hit list, then he will open each document and then perform a search inside the open document. What just happened here is all documents have not only been retrieved on CMOD server, but also downloaded to the ICN and rendered for display. This brings more overhead on the ICN server in addition to the overhead on the CMOD server. And it takes much longer for user to accomplish what he wants to do. Since people are most expensive resources in most cases, I prefer to cater for the human elemnet and let the computer do the work.

My point is, in the grand scheme of things, if we can calibrate the search range to be minimal; and if users are trained very well for how potent and yet dangerous this search is, then it is still cheaper to do it even for large reports. Also, users must be trained to wait for the Full Text Search results to come back and they should never kill a running search and start a new one. Each time they kill their session and start a new Full Text Search, the previous search continues on the CMOD server and keeps running until done.

Good Luck with this approach and please make sure users are very well aware of what they are doing and what really happens in CMOD behind the scenes when using this feature.

-Mehmet

23

Report Indexing / Re: Changing indexed report to non index report

« on: September 01, 2020, 06:08:36 AM »

In addition to Darrell's suggestion of loading the report in full and enabling Full Report Browse capability, maybe you can go the extra mile and add a Full Text Search field to the folder criteria so that hit list is generated only with the reports that have the keyword being searched.

If users don't have Windows client, full report browse will not be available on ICN. ICN doesn't support Full Report Browse. However, Full Text search folder field in ICN will reduce hit list generated and still be very valuable and I strongly recommend it. It is a better approach than a user individually opening each document (or full report) from the hit list and then performing a search for the keyword(s) inside the documents.

24

Report Indexing / Re: Text file autoload

« on: August 24, 2020, 01:08:12 PM »

I think you will need to rename your file to .ARD before you can auto-load with ARSLOAD daemon mode.

Check this link for some detail: https://www.ibm.com/support/knowledgecenter/SSQHWE_10.5.0/com.ibm.ondemand.ir.doc/dodiz045.htm

The relevant section:

Daemon mode
The ARSLOAD program runs as a daemon (UNIX, z/OS and IBM® i servers) or service (Windows servers) to periodically check a specified directory for input files to process. When running ARSLOAD in daemon mode, a dummy file with the file type extension of .ARD is required to initiate a load process. In addition, the Generic indexer parameter file (.ind) must be located in the specified directory. The GROUP_FILENAME: parameter in the .ind file specifies the full path name of the actual input file to be processed.

25

MP Server / Re: Reconciliation via arsdoc query

« on: August 14, 2020, 11:53:34 AM »

In shell, comm command could be a life saver if you have your files sorted and in same format.
You can identify the delta (record in in file 1 but not in file 2; or, in file 2 but not in file 1) between 2 files. 

I compared smaller files and it was very fast for what I needed to do.
I am not sure how it could perform in a file with 200,000,000 records. Maybe you can try a proof of concept with smaller file size and find out for yourself if this is something you could leverage partially into your general solution.

Best of luck,
Mehmet

26

z/OS Server / Has anyone configured z/OS download/arsjesd between 2 mainframe LPARs?

« on: August 11, 2020, 01:36:19 PM »

I believe z/OS Download utility could be used to transfer spooled reports between mainframe LPARs as long as arsjesd is running on the OMVS segment of a receiving LPAR. Am I right?

Is there anyone using this kind of configuration? If so, are there any issues to be aware of?

Best Regards,
Mehmet

27

Content Navigator / Re: ICN Mointoring?

« on: July 24, 2020, 05:26:55 AM »

I am not aware of a script to do the verification. However, if you can make it to deployment with no errors, https://www.ibm.com/support/knowledgecenter/SSEUEX_3.0.7/com.ibm.installingeuc.doc/eucvr000.htm, next you will need to find out which ports ICN runs on and verify the admin desktop.  The entire path that leads to this point could be filled with a lot of places where something could happen to cause an error. If you can make it to this point correctly and if you can log in to admin console, you will need some monitoring tools such as New Relic if  your organization uses it and check its monitoring capabilities.

For example, at my current project, I have set up monitoring for each server's IBM HTTP Server, Websphere Admin Console, ICN Admin Desktop and other custom desktops. This is one way to monitor the ICN server. Additionally, New Relic gives you capabilities to monitor the file system and CPU usage. I have alerts to monitor high CPU usage as well as increased file system use.  Depending on what you want to monitor, these tools are quite capable of providing various intuitive methods to monitor the servers.

I would suggest you start asking your organization's server support teams and find out how they monitor URLs, servers, file systems, server CPUs etc..

Also, please let us know if you found a verification script and also what you did for monitoring and alerting.

Best of luck!

Mehmet

28

z/OS Server / Re: Loading of multiple MVS PS files into CMOD

« on: July 21, 2020, 05:52:18 AM »

Have you analyzed the file sizes? Do they arrive all at once? How many files are exceptionally large and when they arrive they take up most of the space? I have seen in multiple places I worked about less than 10% of the files make up more than 90% of the storage foot print. So, start with analyzing the data and finding the "inflection point" for when the spool could fill up if these files arrive at once. Once largest files are identified, load them via custom JCLs without sending them to  spool at all. Others could still go to spool if you are already set up to process with ARSLOAD started task. You can also talk to your z/OS systems programmers to beef up the spool to max allowed. I have done that in one of the migrations where we used z/OS download via spool to feed a parallel processing non-mainframe system that we were newly building.

Solution sometimes is a mitigation where multiple methods are used and not one is chosen versus the others.

Also, I have seen a lot of times that many files of similar content go into same AG/APP. You need to analyze that as well. If you see clusters of files with very similar naming pattern and content all go into same AG/APP, you can isolate these as well even if they are small files and feed them directly to CMOD with a JCL triggered by arrival of such files. This way, you can reduce number of JCLs you need to set up while bypassing the spool for as many files as you can.

In our business, data drives the process. You must analyze the data and determine best approach with what you find out.

29

Report Indexing / Re: ACIF line Date Indexing Failed

« on: July 14, 2020, 05:39:57 AM »

The only valid carriage control characters that could appear in column 1  can be seen here at the link: https://en.wikipedia.org/wiki/ASA_carriage_control_characters

If the original document was not created with valid Carriage Control characters and you try to process it as if it has valid carriage controls, this error will happen.

Use "arsxline -s" option to sniff the original document properties before modifications and apply them in your indexer.

-Mehmet

30

MP Server / Re: Default location for trace file?

« on: July 02, 2020, 12:07:35 PM »

if you have root access, maybe:
 
find / -type f | grep "trace.log"

....or some variant of this command could help to traverse all directories.