Oracle 7445 Creating EM Repository
zigzagdna
I am using Oracle 11.1.0.7 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.
joel garry
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.
jg
--
@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.
zigzagdna
Thanks. I created a ticket with Oracle more than a week ago which was
assigned to a developer. I have not gotten any response. I did look
into the things you discuss in your e-mail but it appears what is
happening is 11g is different. I also found a bug on same issue on
metalink, so I think it is some Oracle internal issue.