In a prior thread you mentioned - "css has now determined that 534+ is not compatible with win2003 server (its using iis6 and thats the prob). "
Can you be more specific on the issues? - and do you have any idea why its not an advisory on the IBI tech site. We are about to install 534 on a Win2003 server and are very curious of the details. thanks.
big jobs loop indefinitely; its an iis6 problem; so all our etl jobs would never finish and never die. in one of my threads i posted a wee bit of code that CSS gave me to test . the infinite looping was thought , by css, to be an &ECHO problem. but after alot of pressure, they agreed to investigate further and finally agreed that the infinite looping was not just &ECHO related, and they discovered it was an IIS6 problem. so thats where we stand. we killed server2003 and dropped back to win2k, and further , dropped back to 527. You can read the thread, its on about page 3 of this site, has about 70+ comments. If your etl jobs are small, you might be able to get away with 2003. Let us know how you do, ok? As to why its not made public, i wouldn't know.
Is Windows 2003 and Apache Tomcat (instead of IIS6) with WebFOCUS 5.3.4 a viable option?
You mention ETL as the offending product here but I don't see any mention of ETL in the other posts. My question here is, does this affect WebFOCUS reporting or only ETL?.....Kelly
its big jobs. my etl jobs are big. 3000+ lines of code...any big job...read the original thread and copy the testing code and run it...its just a read of the car file within a loop from 1 to 1000 and it will eat up your cpu and never stop.
Here's the case number, as ibi wrote it up;
2005-06-11 11622002 History 0 RDUP:Excessive Logging causes W3WP.EXE to consume CPU time In Product Division 2005-08
I don't know about tomcat, francis, but try and let us know?
Unfortunately, I don't have Windows 2003 server to test with.
I have a Windows XP / WebFOCUS 5.3.3 / Apache Tomcat 5.0.28 environment, no ETL.
Does this looping problem exist in release 7 as well? We're about to put 7.1 on a Windows 2003 Server.
This is a timely message. We are upgrading to 5.3.4 and thought there were no more gotcha's.
I have reproduced your RA (Report Assistant) hang in the MRE, Susannah, but the code to reproduce the infinite loop, from the earlier messages, executes properly in both self-serve and MRE.
I'd like to review the discussion from your case 11622002, if you don't mind, but am unable to get to it. Could you post it, or e-mail it, or let me know if it is possible to get to it via this site?
tglaser,we've dropped back to 527, so i've core-dumped the details and will wait for a long time before trying any other 53n, but i recall that 534 revealed a new incompat, a foc 1517 error on every single fexecution. and its only 2003 server .. an iis6 prob; on win2k, no looping.
We can't get to each other's cases, but your CSS person can get to all the cases. so you can get your css person to look it up for you.
Thank you for the quick response. I've already recommended to my management based on your posts and our testing, stopping the 5.3.4 upgrade pending further evaluation ... and cc-ed our CSS rep for response.
We are new to the Webfocus environment. It is a challenging environment to manage.
As a follow-up; I was running Sun Java version 1.5.0_n on my desktop, and Report Assistant did not instantiate. The "Java Applet Loading ..." never completed. When I rolled Java back to 1.4.2_06, Report Assistant loaded and ran.
We seem to have no RA issues at the present time, or looping issues or foc errors with WF 5.3.4B Gen 818 on a Windows Server 2003 EE with IIS 6, with Java 1.4.2_06 on the Desktop.
t, we bailed on sun java altogther (for exact same and other probs) and are running microsoft jvm, which you can download from ibi support site. i highly recommend that.
Did you run the car file looping test fex on 534/2003? and it didn't gag? if so, 534b gen 818 must have fixed the prob, (but of course IBI doesn't ever tell anyone.)
We've had issues with sun java on MRE 5.3.4 and upper versions past 1.2.4 or however the number it. I've found that keeping JVM as the active has the least problems. We have 7.1 on a LINUX server and it doesn't like the higher levels of sun java.
For the most part our 5.3.4 is running however every once in a while we get a java socket exception out of caster that makes no sense in code that has been running for months and months in caster. I've had to rearrange titles so as not to have headings that have a spot marked value that ends in a '1' (one).
IBI gave a helpful explanation of where the java is active, certified versions.
Yes, I ran the car file looping test fex on 534/2003 and it didn't gag. So something good happened between the version you ran and the version I am running. 534 allows us to put in place better dbms-enforced security on report authorizations, so we'll probably roll this out pretty quickly.
Our corporate standard is Sun Java and that will not change, we'll have to live with whatever conflicts there may be between Sun and IBI. Thanks for the heads up about problems we will likely encounter; this forum is a great tool for that type information.
|Powered by Social Strata|