CDO, Mapi file versions

[Micah_Wyenn]

Hey there folks,
I�m starting to see a lot of instability in the field around EV services after the DST patches get tossed in to environments. I need the official word from Symantec engineering, support, and professional services on what versions of CDO, mapi32.dll, and mapisvc.inf should be on the EV servers in order for them to run without these issues. As you can imagine, this is a fairly hot question, and I�d really appreciate some guidance on this. Thanks!

Micah Wyenn

[Michael_Bilsbor]

Hi,

there should be no changes. What issues are you seeing, can you give us details of the errors.

Deployment scanner should be telling you if there is a problem or not.
Mike

[Micah_Wyenn]

Well,
With 6.0 you used to put the older mapi32.dll in place otherwise it error'd, and replaced the mapisvc.inf's to the latest versions (the ESM). i'm wondering if we should still be doing that with 7.0.
As far as the instability, I'm seeing a bunch of these:


Instruction at "0x7c8396d0" ref'd memory at "0x00000014". Could not be
read

Which are then accompanied by either the journal or archive tasks dyin'. I've also noticed a rash of 3310's, but those are so generic it could be anything.

micah

[Michael_Bilsbor]

V6 sp1 and above and V7 'take care' of mapi and make sure the right one is used (the right one being the one ev needs...) so you shouldn't have to be doing anything like that. If you are you need to say what error you are getting that fiddling with mapi32.dll fixes (as if you're right then our code should be doing more to sort that out)

The errors like that won't be related to MAPi, just code.
I think we fixed some of those problems in V7 sp1.

[Micah_Wyenn]

Kewl,
I'm going to mod the setup procedures I'm using then...thanks!

micah

[Michael_Bilsbor]

ok good news.
We check mapisvc.inf and make do a 'fixmapi' if we have to.
The agents I think will fix mapi if they need to as will admin service if my memory serves me right.