Re: "Cannot obtain globalId for an object which..."
Re: "Cannot obtain globalId for an object which..."
- Subject: Re: "Cannot obtain globalId for an object which..."
- From: email@hidden
- Date: Fri, 14 Nov 2003 16:12:40 CDT
- Priority: 3 (Normal)
[demime could not interpret encoding binary - treating as plain text]
I've been struggling with a related problem in my logs. Like you, it
seems to only happen every once in a while---if I mess with the
(deployed!) software enough, it will happen, but I can't make it happen
on demand. I can't make it happen at all in non-deployed instance, so
this makes me think it has something to do with locking/concurrency. But
I can't figure out what---as far as I can tell, I am locking all my ECs
properly.
My stack trace is different than yours though, so it may not be the same
problem. My problem occurs when I try to delete an EO, and the strack
trace indicates that the exception is generally (perhaps always, not
sure) thrown when propagatingDeleteFromEditingContext, and thrown by
another EC which is receiving this message.
I just deployed my apps with new code in that prints out more
diagnostics, to try and figure out what all the ECs involved are, and
what's going on. If I get enough information to understand the situation
better, I'll post it. It's a very frustrating problem for me currently.
I suspect it's an EOF bug, but I don't really have enough information to
file a bug report---my experience has been that a bug report without a
demonstration case is rarely useful to Apple engineers, and I don't think
they want me to send them my millions of lines of app/framework code, and
say, gee, when you deploy this and enough users start using it, the
problem starts happening, and it _could_ be a bug in my own code although
I suspect an EOF bug.
--Jonathan
On Fri, 14 Nov 2003 14:33:30 +0100 Kaj Hejer wrote:
> At 12:56 -0700 10-06-03, Chuck Hill wrote:
> >Hi Kaj,
> >
> >First off,
> >> I also get a bit confused about what "databaseContext's active
> >> editingContext" mean.
> >
> >When an EC starts to do anything outside of its sandbox (fetching
> >and saving), it locks the database context(s) that it is using.
> >Only one EC can lock a database context at a time. When an EC locks
> >a database context that EC becomes the database context's active
> >editing context.
> >
> >Is it possible that this has been related, in code, to an Pasient in
> >a different editing context?
>
>
> Hi!
>
> Thank again for this info!
>
> Last time I posted about this issue it was on WO5.1.3. Now I have
> this issue on WO5.2.1.
>
> The stacktrace is available at http://folk.uio.no/kajh/tmp/bug20031114.txt
>
> We use only session.defaultEditingContext and childEditingContext of
> this ec. We lock/unlock our childEditingContext in
> session.awake/sleep.
>
> As far as I can see we don't have any relatinships between objects in
> different editingContext.
>
> This issue seems to happen at random (1 of 30 times? I havn't yet any
> reliable data for how often it occur).
>
> I havn't yet testet this on WO5.2.2, but we are goding to do that
> real soon!
>
> Config: WO5.2.1, solaris, java 1.3.1.
>
> Has anybody else here seen anything similar? In case this is a bug,
> does anybody know if this is fixed in WO5.2.2?
>
> I filed a bugreport for this issue (#3484195)
>
>
>
> -Kaj :)
> _______________________________________________
> webobjects-dev mailing list | email@hidden
> Help/Unsubscribe/Archives:
> http://www.lists.apple.com/mailman/listinfo/webobjects-dev
> Do not post admin requests to the list. They will be ignored.
_______________________________________________
webobjects-dev mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/webobjects-dev
Do not post admin requests to the list. They will be ignored.