DSXchange

Good Day

When I tried to LOGTO a project we get the
Read operation failure. >
message. Any idea why this message is shown? Appreciate your support.

From where are you executing the LOGTO command?

Please check permissions on the project from which you are attempting to LOGTO and the project to which you are attempting to transfer.

Finally please check that you have read permission to the UV.ACCOUNT hashed file in the UV account.

We are logging from UNIX. Trying to log to a specific project then we get this message. It was working fine two days back. We have rebuild the indexes successfully.
Here is the error description
From previous run
Read operation failure. DataStage Job 1042 Phantom 14583
[14593] Done : DSD.RUN INITIALIZE_ANB_JOB_STATUS.PYL_STAGING_SEQ_INITIALIZE_ANB_JOB_STATUS 0/0/1/0/0
Program "DSD.WriteLog": Line 265, Read operation failure. Attempting to Cleanup after ABORT raised in stage StagingPayroll_DailySeq..JobControl
Program "DSD.WriteLog": Line 265, Read operation failure.

Look for permissions changes on anything that may have been made in the last couple of days, particularly the DS_RESENU hashed file and the SYS.MESSAGE hashed file and catdir directory. Indeed, search for anything in the project or in the DSEngine directory that has been modified since the job last ran without this error.

Good Day
We tried to use the following command and gets a

Warning: file load discrepancy detected. Below is full extract.


>UVFIXFILE DS_JOBS

Beginning TRACE of DS_JOBS.
TRACE of DS_JOBS completed.

Warning: file load discrepancy detected.

Scanning free buffer chain.
Scan complete.

Scanning overflow buffers.
Scan complete.

48 group(s) processed.
62 group buffer(s) processed.
1016 record(s) processed.
Number of data bytes = 77936.

DS_JOBS load discrepancy can be ignored - it's not fatal, and will correct itself in time (next time the server is shut down). Why did you only run UVFIXFILE on DS_JOBS? Did you check permissions, as I asked?

We have checked the permissions there is no change since the last successfull run.

We have restarted the server, still the problem persists. Here are the extracts.

>LOGTO EDWPROD
Read operation failure. >
>UVFIXFILE DS_JOBS

Beginning TRACE of DS_JOBS.
TRACE of DS_JOBS completed.

Warning: file load discrepancy detected.

Scanning free buffer chain.
Scan complete.

Scanning overflow buffers.
Scan complete.

48 group(s) processed.
62 group buffer(s) processed.
1016 record(s) processed.
Number of data bytes = 77936.


>LOGTO EDWPROD
Read operation failure. >
>UVFIXFILE DS_JOBOBJECTS

Beginning TRACE of DS_JOBOBJECTS.
TRACE of DS_JOBOBJECTS completed.

Warning: file load discrepancy detected.

Scanning free buffer chain.
Scan complete.

Scanning overflow buffers.
Scan complete.

2474 group(s) processed.
3956 group buffer(s) processed.
14876 record(s) processed.
Number of data bytes = 4076596.

Please execute the following commands and post the results. The WHO command determines whether the LOGTO command worked successfully. Examining the LOGIN item in VOC is to try to determine whether the read operation failure is occurring ther.

We have tried your commands. Here are the results.

>LOGTO EDWPROD
Read operation failure. >
>WHO
17 EDWPROD From dsadm
>LIST.ITEM VOC 'LOGIN'

LIST.ITEM VOC "LOGIN" 10:09:36 01-24-09 PAGE 1

LOGIN
001 PA - DataStage
002 IF @TTY NE 'phantom' THEN GO END.OF.LOGIN
003 NOTIFY ON
004 END.OF.LOGIN:

>LOGTO UV
>LOGTO EDWPROD
Read operation failure. >

Some progress has been made. Since LOGTO UV did not generate the error we know that there is no problem with the LOGTO command, so the read operation failure is in the EDWPROD "account" (project). There is no read operation in the automatically executed LOGIN item (which is why I wanted to see it).

Are there other DataStage projects on your server? Can you please try LOGTO any one of those, each with a WHO command, then LOGTO back to EDWPROD ?

Here is the login details on another project

>LOGTO SiebelOnDemand
>who
17 SiebelOnDemand From dsadm
>LIST.ITEM VOC 'LOGIN'

LIST.ITEM VOC "LOGIN" 17:06:22 01-24-09 PAGE 1

LOGIN
001 PA - DataStage
002 IF @TTY NE 'phantom' THEN GO END.OF.LOGIN
003 NOTIFY ON
004 END.OF.LOGIN:

>LOGTO UV
>LOGTO EDWPROD
Read operation failure. >

OK, it's something only in the EDWPROD project that's triggering the read operation failure. Can you please try these variants of commands from the LOGIN paragraph, again posting the result?

We have tried your commands here are the results.

>IF @TTY NE 'phantom' THEN DISPLAY Interactive
>NOTIFY ON
>LOGTO EDWPROD
Read operation failure. >
>IF @TTY NE 'phantom' THEN DISPLAY Interactive
>NOTIFY ON

What we could observe is that when we try to logto any project
VOC file under this project is been accessed. We presumed that this VOC file has got some issues, we checked the permissions there is no issues and alternatively we tried the following commands.

$ fixtool -fix -file VOC
Filename: VOC

Start Diagnostics.
0 Errors found in physical structure.
0 Errors found in file header.
0 Errors found in file groups.
0 Other errors found in file.

7465 records processed.

No errors were found.


Fixtool Completed.

It's frustrating how slowly we're progressing, I know, but we have to eliminate the impossible (Sherlock Holmes).

The only other local "file" that gets read when an account is accessed is &SAVEDLISTS&. Even so, please check the permissions on VOCLIB, D_VOC, D_VOCLIB, D_&SAVEDLISTS& and &SAVEDLISTS&. Please also verify whether you can execute the command COUNT &SAVEDLISTS&.

(Later I may ask you to CLEAR.FILE &SAVEDLISTS& but don't do that just yet.)

Yes, its frustrating.

Here are the results.

>COUNT &SAVEDLISTS&

28 records counted.
>



-rwxrwxr-x 1 dsadm dstage 16384 Aug 21 2007 VOCLIB
-rwxrwxr-x 1 dsadm dstage 4096 Dec 31 2007 D_VOC
-rwxrwxr-x 1 dsadm dstage 2048 Aug 21 2007 D_VOCLIB