Hi Anthony, Thanks for the response. I'm not aware of any specific change made in the last several years regarding this behaviour but will do some research to see if I can find if/when this may have been changed.
However, if a specific change was made, it would seem like it was long enough (i.e. release 10.0 or 9.6?) ago that it would not have much of an impact on Release 10.x upgrade plans at this point in time. Thanks,, Neil Neil Morris SWG Client Support - Information Management Software UniVerse ATS b IBM U2 Client Support Team 1-877-676-9452 [email protected] "Anthony W. Youngman" <[email protected] To on.co.uk> [email protected] Sent by: cc owner-u2-us...@li Wally Terhune/Denver/i...@ibmus, stserver.u2ug.org Neil Morris/Portland/i...@ibmus Subject Re: [U2] Does anybody have a hot 03/21/2009 06:06 backup server? {Unclassified} AM Please respond to u2-us...@listserv er.u2ug.org In message <of087fe71c.cb64f47e-on8525757f.0077b5a9-8525757f.0078f...@us.ibm.com>, Neil Morris <[email protected]> writes >Hi Mike, > >Your inquiry was brought to my attention today. > >Could you possibly clarify what is meant by "headers are flushed when they >are updated" in regard to UniVerse dynamic files? Is this in reference to >the information in the T30FILE structure in the UniVerse disk shared memory >segment being written to the header of the file when an update would result >in changing one of the dynamic file parameters stored in the header? Or >does this comment refer to the updated information in the header of the >file being flushed to the physical disk via a sync operation? > >If the former, then it is my understanding that UniVerse has always updated >the header immediately when a dynamic file parameter changed and there is >no specific release in which this capability was added. If the latter, the >only way to force the update to be flushed to disk immediately is to have >the file activated for logging with UniVerse Transaction Logging and have >Checkpoint mode enabled. > >Perhaps I'm not understanding the question so any clarification you can >provide would be appreciated. > Hi Neil, It's long been "common knowledge" (which doesn't mean it's correct :-) that UV read the dynamic file parameters into memory when the file was opened, and didn't always write changed information back until the file was closed. You were therefore advised that it was not safe to copy even a quiescent dynamic file at OS level, because the header and data might not be in sync. I remember a posting (from an IBM'er) saying that this had changed round about rev 10, and that header and data should now always be in sync on disk. (ie the post was about rev 10, the change might have been earlier. But the implication was that it *had* been changed.) Cheers, Wol -- Anthony W. Youngman <[email protected]> 'Yings, yow graley yin! Suz ae rikt dheu,' said the blue man, taking the thimble. 'What *is* he?' said Magrat. 'They're gnomes,' said Nanny. The man lowered the thimble. 'Pictsies!' Carpe Jugulum, Terry Pratchett 1998 Visit the MaVerick web-site - <http://www.maverick-dbms.org> Open Source Pick ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/ [demime 1.01d removed an attachment of type image/gif which had a name of graycol.gif] [demime 1.01d removed an attachment of type image/gif which had a name of pic24619.gif] [demime 1.01d removed an attachment of type image/gif which had a name of ecblank.gif] ------- u2-users mailing list [email protected] To unsubscribe please visit http://listserver.u2ug.org/
