Re: File system error occurred. Error number 3029.

[Scott Klement <midrange-l@xxxxxxxxxxxxxxxx>]

Hi Gary,

> Yesterday, the process failed repeatedly, receiving "Error number
> 3029." when attempting to process a file on the IFS using the
> QP0L_ProcessSubtree api (thank you to Scott Klement).

No problem!  Any time you want a process to fail with 3029, I'm your man!


> The confusing thing to me is why the resource busy message would prevent
> completion of step 1), but allow the "busy resource" to be moved in step
> 2).

To me, it makes sense that an object can be moved when it's in use, but 
can't be changed when it's in use.  After all, you can move an IFS 
object without affecting a program that has it open.  (Heck, you can 
even delete an IFS object without affecting a program that has it open!)

But...  I don't understand why your process doesn't abort when it gets 
the EBUSY error?  If step 1 fails, why does your program continue on to 
step 2?  Perhaps that's a bug.


> I have reviewed midrange list discussions on CPE3029/resource busy, but,
> as stated above,
> am not certain if the busy resource is the file object or the
> directory/folder.
>
> I welcome suggestions on how to research, analyze or avoid this problem.

When you open a file with the open() API, you can specify O_SHARE_NONE, 
O_SHARE_RDONLY, O_SHARE_WRONLY, or O_SHARE_RDWR.

Perhaps your program should try to (at least momentarily) open the file 
and specify O_SHARE_NONE, to make sure you're the only one using an 
object before trying to manipulate it? I'm not familiar enough with the 
process to know if I'm right about that -- but it's something to consider.