Re: Oracle 7445 Creating EM Repository
Date: Wed, 3 Feb 2010 09:13:34 -0800 (PST)
On Feb 2, 8:26 pm, zigzagdna <zigzag..._at_yahoo.com> wrote:
> I am using Oracle 126.96.36.199 on HP UNIX 11.23 ia 64. When creating em
> repository for database control , I get ORA-7445 error. When I created
> my first database instance I was lucky to get EM repository. IN my
> other db instances I consistently get ORA-7445 error with core dump.
> Does any one have work around for it? DB Control is so useful and alas
> I cannot create it. Setting up Grid Control is lot of work and who
> knows I may run into same problem.
Work with support, they seem to need some prodding to fix things in the hp space, especially for dbconsole. In 10204 it is just lame, I'm going through that now with support, and seeing things like "Couldn't duplicate" on the most obvious screwups. If what I'm seeing has any relation at all to 11g, check to see that the emctl.pl actually shuts and starts the proper java session when you give it commands that involve that. Check to see that the logs don't mention anything about ports in use (um, assuming you are using different ports for each?). It's supposed to be fixed in 11g, but what I'm seeing is that emctl doesn't shut down the java session that actually controls the port, so it sortakinda works the second time, but really doesn't, so you get into strange mixed states, sometimes the only way to tell is to check the date of the java process. This comes from the code in emctl.pl being for other platforms, so I can't help wondering from your post if the "fix" for 11g is just to make it die rather than being confused. I haven't got to look at that pl yet so I don't know how it distinguishes between the dbconsole process and the java process (which is a semantic screwup as far as I can tell in the 10g). My support analyst clarified that:
"3 processes are started with the DBConsole application.
-- emagent.exe -- java.exe -- perl.exe"
Funny, I don't see any exe's on hp-ux. Anyways, I also noticed some crashes and restarts of java itself (evidence trail starts in emdb.nohup), so that spawned another SR, and I still don't know what is going on. But I was able to determine, be sure all emagent, java and perl related processes are extinguished when you do any kind of reconfiguration. That's easy for me, as only dbconsole is using java, but of course, YMMV. The hard part for me is remembering that all the instructions on reconfiguring, patching and securing dbconsole from MOS and everywhere else don't account for this (unix platforms that aren't linux) problem, and so tend to make any problem worse.
In your instances getting 7445, does anything else cause that, or is it only dbconsole? There are MOS documents on what to do about 7445 in general, I'm sure.
-- _at_home.com is bogus. "Give a cat a fish, he eats for a day. Try to teach a cat to fish, and he will claw your face until you give him a fish." - my kid told me that.Received on Wed Feb 03 2010 - 11:13:34 CST