Each site has it’s own recovery and error databases.
If you run hcidbdump make sure you are in the correct site. From the GUI look at the statusbar at the bottom. From the command line run ‘showroot’ before hcidbdump.
That is what I thought. When I am in the production site and run the hcidump around half of the entries are for the production site as expected, the other half are from a different site.
Should I try shutting down all the threads in each site and doing a database initialization for each site to try to correct this problem.
Wow. any chance the database files were copied/promoted from one site to another?
If the data from the “other site” is old this may be what happened. You can check the date/time of the transactions with hcidbdump -r -l
If you’re OK with blowing away your databases you can go ahead and stop all sites and do “hcidbinit -A -C” on each site. Note this will wipe all data in recovery and error databases.
I know Ben’s on leave this week but it might be worth mentioning it to him when he’s back. Although there’s not a lot he can do as I can’t think of any reason why it should be doing that.
I’d go for the hcidbinit -AC if you can.
If you want to keep in touch my email is garry.fisher at adaro.net
Thanks for the input. I took the dynamite option and initialised the databases. Which seems to have cured the problem for now.
Its a strange one this as there had been no changes as far as I am aware off. We have a couple of maintenance tasks that are run from cron jobs which do set various qdx environment variables and this is probably going to be the first place i look.
I have found another problem whilst looking into that one, the debug output from one thread is now appearing in the log of another thread, at least it is within the same site I’m beyond baffled 🙄
Garry, I will give Ben a call next week. hope things are going well in your new job!
Jason
Author
Replies
Viewing 4 reply threads
The forum ‘Cloverleaf’ is closed to new topics and replies.
.
.