Re: Re: finding inplace alters
Hi Art, the idea is walking all (or certain) dbspaces' tblspace tblspaces, relate partition header pages carrying a version pointer to their secondary partition pages (version pages + dependent alter descriptor pages), in correct order, then walk the contained alter descriptors and read out the versions' page counters and relate them to the partition's name and overall page count. I don't think this would be feasible using (existing) SMI tables (unless you'd want to extract single bytes by offsets which would require intimate structure knowledge). I'd be in favour of having new SMI tables for this purpose though these probably wouldn't be trivial to implement. They'd also probably not walk tblspace tblspcace... (as soon as a partition is loaded, all its IPA information is availble in memory). One could also imagine other dedicated server/sysadmin functionality in this direction ... not only to find and show all sorts of IPA informations, but also to have a more selective way of removing outstanding IPAs (e.g. dummy updates on only as many rows per old-version page as are required to update the whole page). On the other hand in-place alters shouldn't be such an issue any more nowadays, right? Though you should of course have a way to know what state your tables are in. Regards, Andreas From: Art Kagel <art.kagel@gmail.com> To: Andreas Legner/Germany/IBM@IBMDE Cc: Davorin Kremenjas <davorin.kremenjas@gmail.com>, informix-list@iiug.org Date: 20.12.2010 18:48 Subject: Re: Re: finding inplace alters Andreas, you could probably rewrite it to use the sysmaster:syspaghdr table and then you may be able to release the source. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) IIUG Board of Directors (art@iiug.org) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Mon, Dec 20, 2010 at 12:41 PM, Andreas Legner < andreas.legner@de.ibm.com> wrote: I could offer a tool called 'onion' that, among other things, can provide the desired informations directly from versions' page counters on (secondary) partition pages, so no data and other pages need to be scanned and you can get what you want in a second or two. The engine will not sense a bit since the tool is reading directly from disks. Nothing official, designed for support purposes, but usually serving its purpose... Let me know if you need it (platform required since it is a binary). Andreas On 20:59, Davorin Kremenjas wrote: > On Nov 18, 12:21 am, Fernando Nunes <domusonl...@gmail.com> wrote: > >> AFAIK there is no quick way to find pending inplace alters... Only a slow >> process. Not sure if Panther changed this... > > There is a good document somewhere on the IBM web and it is slow, but > it works, returns only tables with pending IPAs. > Since it's usually the biggest tables which end up with pending IPAs > (there's never enough time/resources to get red of them properly) I > was wondering if this could be further narrowed down to specific pages > in those tables. Never got to the bottom of this though. > > Fernando, would you know if page flags could provide an answer? For > example, sysmaster:syspaghdr.pg_flags. > > Thanks > > Davorin > _______________________________________________ Informix-list mailing list Informix-list@iiug.org http://www.iiug.org/mailman/listinfo/informix-list