Re: Jobs do not run

From: Martin T. <0xCDCDCDCD_at_gmx.at>
Date: Wed, 08 Aug 2007 12:42:46 +0200
Message-ID: <46b99eb6$0$1348$834e42db@reader.greatnowhere.com>

jonathan.saltiel_at_gmail.com wrote:
> On Aug 6, 1:55 am, "Martin T." <0xCDCDC..._at_gmx.at> wrote:

>> sybra..._at_hccnet.nl wrote:
>>> Comments embedded
>>> On Fri, 03 Aug 2007 17:30:51 +0200, "Martin T." <0xCDCDC..._at_gmx.at>
>>> wrote:
>>>> The problem only exists in our test db instance (at the moment)
>>>> I fully expected you and Daniel to tell me that. (Given the fact that
>>>> there are 7 patch sets we haven't even looked at it might even have some
>>>> foundation.)
>>> All patch sets are cumulative
>>>> Let's for a moment assume we were running Oracle 10g2 with the newest
>>>> patch and hit this. What would you propose except calling Oracle?
>>> a) The most common cause for jobs not running is forgetting to commit
>> I know this. That you replied it to my OP just goes to show that you did
>> only read "jobs" and "...0.1" and then shouted something at me.
>>
>>> b) the problem doesn't exist in 9.2.0.8
>> Interesting ... which crystal ball told you that?
>> I tell you what: it also does not exist in 9.2.0.1 on 40 other machines
>> with the same HW and SW setup. It also did not exist on this machine
>> prior to one week ago.
>>
>> Well whatever.
>>
>> regards,
>> Martin

>
> Some DBAs on this list need a reality check.
>
> In the real world Databases don't get upgraded at the drop of a hat -
> New patches sometimes introduce new issues.
> Upgrading takes lots of planning and testing, and based on the case
> mentioned I would not upgrade the database on the count that it might
> solve the issue, especially since it is not an issue anywhere else and
> not prior to today.
>
> Once you understand the root cause of the current issue, and you find
> out for certain that an upgrade will fix it (by testing elsewhere),
> then you need to upgrade.
>
> Jonathan
>

To wrap this up:

What we did was the following (remember this is a test machine) ...

1.) We exported the relevant Schema(s) data
2.) Scratched instance (not Oracle binaries, only database file&config)
3.) Recreated instance, System tablespaces etc. (Which incidentally 

provided a good opportunity to check the creation scripts again) 4.) Imported the Schema(s) data.

Voila: Jobs run again.

I wouldn't go so far as to call it a beautiful solution, but it certainly was less hassle than patching (which may or may not have helped).

best regards,
Martin Received on Wed Aug 08 2007 - 05:42:46 CDT