Hi,
we recently moved from Data Services 3.2 SP to 4.2. SP3 in our development enviroment to check for issues, changes etc. before moving to production. After a series of different problems (some involving internal firewalls, others problems with the repositories..) we are now almost there.
We can execute every job, can change every dataflow etc, but we cannot validate the complex jobs. To ease handling for external scheduling we have 2 large jobs, each consisting of around 100 Workflows, each consisting of 1-2 Workflows. In the old DEV-enviroment and the still existing old-school PROD-enviroment we can validate those, in the migrated DEV-enviroment we get something like this:
In addition we have to close the Designer and reopen it to continue working.
We know the message in cases of memory issues, but as far as we can see, the max used memory is about 1GB, same as in the old enviroment. Designer trace (activated in DSUserConfig.txt) generates no messages. Errorlog.txt gives some hints:
(14.2) 11-03-14 14:34:24 (E) (3688:4704) RES-020107: Referencing an undeclared variable <$JB_ERROR_COUNT>. Declare the variable in the context that it is being used.
# hint from me: variable is declared and reachable
(14.2) 11-03-14 14:36:42 (E) (3688:4704) PAR-010102: Syntax error at line <3>: <>: found <[end of text]> expecting <'(', &ERROR, &ERROR_CONDITION, __AL_LOOKUPEX_TRAN, __AL_SEARCH_REPLACE_TRAN, __RFC_FUNCTION, __AL_SCRIPT_FUNCTION, __AL_STORED_PROCEDURE, __AL_EXTRACT_FROM_XML,__AL_EXTRACT_FROM_JSON, __AL_TRAN_FUNCTION, +, AL_UNSPECIFIED_PARAM, CONVERT, a float, identifier, an integer, a null,PUSHDOWN_SQL, a quoted identifier, ;, a string, a decimal, VARCHAR, VARIABLE, -, +, NOT>. 1 error(s), 0 warning(s).
# hint from me: individual validation does not generate this one. So maybe it is part of the issue.
But validating each Workflow on its own generates no messages..
We can generate the same error message (screenshot above) by validating almost all individual workflows in quick succession. Feels like hitting some kind of limit or timeout we are not aware of.
For the evironment:
- Job server is AIX 6 with 32GB RAM and almost empty disks (but as far as we understand it is not involved in validating jobs)
- Client machine is Win 7, 4GB RAM, also almost empty disks, 64-bit client is installed (14.2.3.549)
- Network connection seems perfectly fine, firewall exeptions for all needed ports are in place
Any idea what is happening? We are afraid to move to production while there is still an unexplained issue.
Thank you in advance
Daniel