Nope, sorry. Why do you ask?

Post #380,107 by CRConrad 9/1/13 6:06:30 PM Reply	Nope, sorry. Why do you ask? Mostly Oracle (and old stuff at that; OWB/Disco rather than ODI/OBIEE), some Cognos and Informatica, more recently smatterings of MS and SAS (back to good-ole DATA STEP, PROC this-and-that, where I began my "career"[] twenty-odd years ago). Oh, and I took a course in IBM Data Manager once; too bad I never got the opportunity to use it for real, because it was quite interesting from a technical viewpoint. But I assume Business Objects's ETL tool-thingy is at least superficially like all the others; all the others certainly are at least superficially all pretty much like each other[+]... So, the height of arrogance: If you want to know something about how it works in practical use, I'll try to answer anyway. Ask away! :-) --- []: To use a hugely over-grand term; I'm not sure what exactly makes one's succession of jobs a "career", but it sure don't feel like I've got that. Some people have careers, I have a job. When I do. [+]: Except perhaps for the aforementioned IBM Data Manager, which differed from the pack in one major technical aspect. -- Christian R. Conrad Same old username (as above), but now on iki.fi (Yeah, yeah, it redirects to the same old GMail... But just in case I ever want to change.)
Post #380,145 by mmoffitt 9/2/13 1:14:25 PM Reply	I had a specific question about Data Services. But your background might be sufficient for you to answer my question (and perhaps confirm my suspicions). We use PostgreSQL Plus Advanced Server from EnterpriseDB. If you're unfamiliar with that, it's a commercial product that has an "Oracle Compatibility Layer" on top of PostgreSQL. Another group where I work uses SAP Data Services to populate some databases they've created. The thing is, when these data flows start, they immediately open about 25 connections each and since the "job" contains many of these flows, they quickly eat all the connections available to the PostgreSQL instance. I'm told that SAP has confirmed that Data Services does "a poor job of handling connections" and having connections opened, used and then left open is not unexpected. We tried using PgBouncer to limit how many connections Data Services uses, but the jobs refused to start (we set the pool as high as 50 and apparently Data Services needs around 120 connections for these jobs, even though their use appears to be sequential and the product will not close a connection after its finished with it - so Data Services just hangs if you limit connections either via a connection pooler or in the database instance itself). SAP/BO's support for PostgreSQL is lackluster at best. I know the reporting service (what used to be BO Enterprise, Infoview, etc. but its named something else now) does not fully support PostgreSQL and a VP there shortly before the SAP acquisition told us in a telephone conference to use MS Sql Server. But our CIO will not switch to that product and will simultaneously not switch to another BI tool - or ETL tool for that matter. I don't know if Data Services' problem with connection handling is limited to PostgreSQL, but I think it must be because I cannot imagine anyone would use the product if even simple ETL flows required hundreds of simultaneous connections. I'm also not convinced that the Data Services authors haven't gotten something completely wrong. I posted here because I thought if you had experience with these tools, you might know something about the purported "connection handling problems". Chiefly, I'm curious to know if we're having this problem because... 1) Our authors don't know what they're doing 2) We're using a backend database that SAP supports only marginally 3) Data Services really does require hundreds of concurrent connections for even simple tasks. Or a combination of all of those. If you have an opinion on it, I'd be interested in hearing it. Thanks in advance.
Post #380,294 by CRConrad 9/5/13 1:13:11 PM Reply	Ah. Sorry, no, I don't really know. But at least I wouldn't go assuming 1; seems more likely to be some combination of 2 and/or 3; both ring true, so I can't say how much of which. -- Christian R. Conrad Same old username (as above), but now on iki.fi (Yeah, yeah, it redirects to the same old GMail... But just in case I ever want to change.)
Post #380,302 by mmoffitt 9/5/13 3:35:03 PM Reply	Thanks. I think its a combination of all three myself, heavily weighted by (2). And, unfortunately, for entirely non-technical reasons nothing can be done about that. ;0( Aside: my group just installed Pentaho CE and has been horsing around with it a little. It seems to work much better with PostgreSQL than the SAP/BO Suite does. Thanks for the input.
Post #418,493 by CRConrad 6/12/17 8:08:14 AM 6/12/17 8:08:14 AM Reply	Came across this while doing a bit of catch-up reading... ...and whaddayaknow, only a bit over a year after that, when starting my current job, I also got to get acquainted with Pentaho. Mostly PDI, the ETL tool, but also a bit of Pentaho BA/"Mondrian", the reporting end. Paid-for version up until three weeks ago, but recently I'm testing the Community Edition on one of our test servers and soon on my own workstation. So I guess the tables are turned: If I have any questions, I can turn to you, Mike (you old Chekist, you), since you must be an old hand at this stuff by now? -- Christian R. Conrad Same old username (as above), but now on iki.fi (Yeah, yeah, it redirects to the same old GMail... But just in case I ever want to change.)
Post #418,495 by mmoffitt 6/12/17 8:19:09 AM 6/12/17 8:19:09 AM Reply	I'll help if able. ;0) bcnu, Mikem It's mourning in America again.

Welcome to IWETHEY!