Problems with Server Groups and Data Services Web Service Calls

[former_member467822]

We have two job servers set up in a server group and we're using web services to execute jobs and then retrieve their log files.

The executing works fine - if one server is down the job will be executed on the other server in the group.

However, if the job is executed on a server other than the one receiving the web service request, then the log files can not be retrieved.

We are hoping that having both job servers store their log files on a shared folder will solve this problem (see previous post) but we're not sure.

I'd like to know if this solution will work - or if there is another recommended solution.

[Former Member]

Hi Manoj

Thanks for confirming the bug. Yes the job server and web server are on the same server.

Dinesh

[Former Member]

Hi Manoj

The webservices.log file has a bunch of lines with the same error-'Unable to collect real-time service definitions'. We are not using the realtime jobs and have not setup the access server.

I removed the contents of the trace file and put only "test", I am able to see the results from both the S1, S2 server admin console batch job page.

When I call the webservice from SOAP UI or FIM, I get the messge "The Trace File does not exist". I have opened a case for this but so far not much help. If you are interested I can set up a webex session for you.

Thanks

Dinesh

[Former Member]

found the issue, this is a bug in XI 3.1, and is fixed in XI 3.2

is your Job server from which you are able to open the log file and Data Serivces Web Server on the same machine ?

the problem is the web server instead of getting the log file from the job server is trying to check if the file exists locally if not , is throwing file doesn't exists error

since you have filed a case already, have that escalated to engineering

[Former Member]

I ran a webservice call for a DS batch job from FIM 7.0. The DS on S1 shut down. The job ran on JS on S2.

Following are the results in the ALVW_HISTORY view

OBJECT_KEY -- 94 (correct value)

INST_SERVER --- job server (correct value)

INST_SERVER_PORT --- job server port (correct value)

NORM_SERVICE --- Job name (correct value)

SERVER_GROUP_NAME (correct value)

ERROR_LOG (correct location but this location is now on S2, since S1 is shut down)

TRACE_LOG (correct location but this location is now on S2, since S1 is shut down)

MONITOR_LOG (correct location but this location is now on S2, since S1 is shut down)

In FIM I get the response 'The Trace file does not exist'.

I tried the same steps from SOAP UI and the results are the same.

I logged into the admin console and I am able to see the trace, log files

Thanks

Dinesh

[Former Member]

can you check for any errors in %LINK_DIR%/log/webservice.log file ?

this is strange if your ERROR_LOG, TRACE_LOG and MONITOR_LOG point to correct location, and you are able to open the files from history page, then it should work from web service call also

there is one more issue in 12.1 with trace, monitor and error file but in that case no output is retured, the problem happens if you have & (ampersand) character some where in the error message printed the error file, take a backup of the trace file and change the content to "test" or some text and then try to open

[Former Member]

Hi Manoj

I am working with Joe on this and I can provide more information. Following is our setup and observation. We are designing this for High Availability of FIM 7.0 with DS 3.1.1.1. FIM 7.0 is a web application that uses DS webservice calls to execute DS batch jobs and publishes the status, traces, log files information to the submitting user within FIM.

- We have 2 servers - S1, S2

- Both S1 and S2 have FIM and DS installated.

- One Local repository is assigned to one JS on each server

- A server group is setup for this repository.

- When both the DS servers are up and running, from S1 a webservice call is made to execute a batch job, the job completes successfully and FIM is able to retireve the trace files to be used for display to the user within the FIM application

- When the DS on S1 is shut down and the FIM application on S1 makes a webservice call to execute a batch job, the job gets executed on the JS on S2, the trace and log files are created under the S2 - JS log file location, but when trying to view the files within FIM application, I get the message - "The Trace log file does not exist".

I tested this setup using SOAP UI and the behaviour is the same. What is interesting is, when I log into the admin console on either S1 or S2, I can see the trace and log files from the job server on which the job was executed.

This definitely looks like a bug. Can u confirm it and also are there any workarounds, like pointing the both the job servers to write logs to the same location. If so, how do we set it up. Is it something supported by SAP BO.

thanks

Dinesh

[Former Member]

I tired the same scenario on DS 3.2, but don't see the issue, let me try this on 12.1.1.1 the same version that you are using

what is the runid that you are passing to the get_trace/error/monitor_file() service ?

the jobserver, port, trace ,montior and log file path are stored in history table, and web service uses the runid (OBJECT_KEY column in ALVW_HISTORY) to get that info from repo and request the job server to get the log files

run a query against ALVW_HISTORY WHERE OBJECT_KEY = runid that you are passing to web service and see the values of the following columns

OBJECT_KEY

INST_SERVER --- job server

INST_SERVER_PORT --- job server port

NORM_SERVICE --- this is the name of the job

SERVER_GROUP_NAME

ERROR_LOG

TRACE_LOG

MONITOR_LOG

is the info correct ?

[Former Member]

if the Get_Error_Log web service is not able to return the error log file for job, in case you are executing job using server group then I think its a bug, you don't have to change your job log directory

are you able to get the trace and monitor file ?

is your job failing with error ? if there are no errors, then the error file will be empty

can you give me the steps that your doing

what is the web service that you are calling to execute the batch job, Run_Batch_Job?

what is the value you are passing in Job server and Server group input parameter ?

I can check if this a bug