Next backup of DBspace rootdbs must be level-0
A user on IDS 9.40 saw "Next backup of DBspace rootdbs must be level-0" in the online log (caused by the level-0 archive timestamp drifting out of range) and wanted his backup script to detect this programmatically rather than by parsing online.log. Art Kagel noted a level-0 archive clears the condition. John Miller supplied a query against sysmaster:sysdbstab testing flags bit 0x20000 (sum/decode) to report whether incrementals are allowed; since bitand didn't exist on that older server, he suggested the sysmaster:bitval() procedure instead. No confirmation from the poster is recorded.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Hi over there, I've found the following warning in my 9.40.FC9 IDS dataserver online.log : "Next backup of DBspace rootdbs must be level-0 backup" The paper "https://www-304.ibm.com/support/docview.wss?uid=swg21429486"; explains that this is due to a timestamp used by IDS to identify related events such as backup levels and that : "Incremental archives (level 1 and 2 archives) depend on the timestamp of the most recent level-0 archive to know which pages need to be archived. If that timestamp goes out of interval (more than half way around timestamp cycle), the incremental archives may miss pages that should pick up for archive or get the pages that should not. To prevent any such archive inconsistency IDS forced user to take a level-0 archive". I know the stamp0, stamp1, stamp2 sysmaster:sysdbstab columns contain different backups timestamp values Can someone explain how to know the level 0 backup timestamp has gone out of interval apart from checking the online.log file ? Thanks.
Certain fuzzy checkpoint operations will put the server on this state. Just take a level 0 archive and all will be well. Art On Jan 21, 2011 9:47 AM, "GILLES TCHAPPI" <ntgilfr@voila.fr> wrote: > Hi over there, > > I've found the following warning in my 9.40.FC9 IDS dataserver online.log : > "Next backup of DBspace rootdbs must be level-0 backup" > > The paper "https://www-304.ibm.com/support/docview.wss?uid=swg21429486"; > explains that this is due to a timestamp used by IDS to identify related > events such as backup levels and that : > > "Incremental archives (level 1 and 2 archives) depend on the timestamp of the > most recent level-0 archive to know which pages need to be archived. If that > timestamp goes out of interval (more than half way around timestamp cycle), > the incremental archives may miss pages that should pick up for archive or get > the pages that should not. To prevent any such archive inconsistency IDS > forced user to take a level-0 archive". > > I know the stamp0, stamp1, stamp2 sysmaster:sysdbstab columns contain > different backups timestamp values > > Can someone explain how to know the level 0 backup timestamp has gone out of > interval apart from checking the online.log file ? > > Thanks. > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > --20cf30433ed2667b40049a5ececf
Thanks for your answer. I know I can take a level 0 backup to resolve this problem. The issue here is that I'm planning to take level 0 backup on sunday and wenesday and level 1 backup the other days and I want my level 1 backup script to : without checking the online.log file, automatically detect if the timestamp has gone out of interval by querrying a sys table for instance and if the timestamp is out of intervall, take level 0 backup otherwise take level 1 backup So, my question was : is'it possible to know the timestamp is out of interval wtih a querry in a sys table Regards
the following select statement will tell you if you can take level 1 or 2 backups. select decode( sum( bitand(flags,'0x20000') ),0,"Incrementals allowed","Incrementals NOT allowed") from sysmaster:sysdbstab; John F. Miller III STSM, Embedability Architect miller3@us.ibm.com 503-578-5645 IBM Informix Dynamic Server (IDS) ids-bounces@iiug.org wrote on 01/22/2011 05:27:01 AM: > From: > > "GILLES TCHAPPI" <ntgilfr@voila.fr> > > To: > > ids@iiug.org > > Date: > > 01/22/2011 05:27 AM > > Subject: > > Re: Next backup of DBspace rootdbs must be level-0 [22531] > > Sent by: > > ids-bounces@iiug.org > > Thanks for your answer. > > I know I can take a level 0 backup to resolve this problem. > > The issue here is that I'm planning to take level 0 backup on > sunday and wenesday and level 1 backup the other days and I want > my level 1 backup script to : > > without checking the online.log file, automatically detect > if the timestamp has gone out of interval by querrying a sys > table for instance and if the timestamp is out of intervall, > take level 0 backup otherwise take level 1 backup > > So, my question was : is'it possible to know the timestamp is > out of interval wtih a querry in a sys table > > Regards > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >
Hi, I've tried the statement but I got the following message " 674: Routine (bitand) can not be resolved." Regards
If bitand is not present then you are running an older version of the server. You can replace it with a stored procedure called sysmaster:bitval() John F. Miller III STSM, Embedability Architect miller3@us.ibm.com 503-578-5645 IBM Informix Dynamic Server (IDS) ids-bounces@iiug.org wrote on 01/24/2011 12:50:12 AM: > From: > > "GILLES TCHAPPI" <ntgilfr@voila.fr> > > To: > > ids@iiug.org > > Date: > > 01/24/2011 12:51 AM > > Subject: > > Re: Next backup of DBspace rootdbs must be level-0 [22543] > > Sent by: > > ids-bounces@iiug.org > > Hi, > > I've tried the statement but I got the following message " 674: Routine > (bitand) can not be resolved." > Regards > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >