Bug 21837 did not require a change to Legato.

The change was completely in OnTap. The question was semantics ctime is the inode change time. [i.e. changes associated with the inode should update the ctime field.] The access time is also part of the inode. Therefore in the strict sense of things, ctime should be updated when the access time is updated.

However, backup programs are the only ones likely to explicitly change access time and for those types of applications updating the inode change time doesn't make much sense. Backup programs usually try to set all the times explicitly.

Greg is pretty new (and a manager) and hopefully you can forgive him. :)

-----Original Message----- From: Matt Phelps [mailto:mphelps@cfa.harvard.edu] Sent: Wednesday, June 07, 2000 6:07 AM To: Linn, Greg Cc: cet1@cus.cam.ac.uk; toasters@mathworks.com Subject: Re: BudTool incremental backups behaving differently when using NDMP

Greg.Linn@netapp.com wrote:

I believe that Matt's problem is different from bug ID 21837.

21837 was identified during investigation of a Legato Networker problem where ctime was being updated inappropriately. The net effect of this was that

files

that hadn't changed were being processed as if they had. I believe that

Matt's

problem is the opposite; files that have changed are not being picked up by

NDMP

incremental dumps. Matt's problem is currently being investigated.

That's right. Greg (or anyone), can/has someone confirmed this behavior? I'm concerned about missing something obvious.

FYI: the resolution for Bug ID 21837 required a change to Networker as well as Data OnTap.

I'm hopeful we won't need a change to BudTool, those are hard to come by these days ;-)