Ontape backup

Hi All, We are running informix 7.13 on HP-UNIX 10.01. We are taking a weekly level 0 backup of informix and writing backup to EMC clarrion. We noticed every time backup will complete at different time, usually it took 14 to 15 hrs to complete. We have a database of size 36 GB. Is that the normal behavior of ontape to backup 36Gb of data in 14 to 15hrs or we can do something to complete it faster? Secondly, can anybody help me to troubleshoot why the backup time varies every time. Regards Deepak

Backup times will depend on the IO throughput of the subsystem to which it is being written and of that of the chunks that are being backed up. Also if the chunks are on some of the same physical structures that the archive is being written to, there will be contention for the disk heads. Finally the configuration of the EMC RAID group that you are writing the archive file to will affect the back up throughput (ie RAID5 is 40% slower than RAID10 for sustained sequential writes which is what the archive is doing). Differences in archive times day to day can be due to activity on the server - more updates/inserts/deletes will require that more pre-image data from the physical log be copied out to the archive after each dbspace's back up is completed to maintain consistency of the archive. Another cause can be contention with other applications for IO bandwidth to the EMC array or even between the machine that is hosting the IDS instance and other machines on the network that may be sharing the same EMC array. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) IIUG Board of Directors (art@iiug.org) See you at the 2010 IIUG Informix Conference April 25-28, 2010 Overland Park (Kansas City), KS www.iiug.org/conf 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, Mar 22, 2010 at 12:50 PM, DEEPAK JOSHI <djoshih@hotmail.com> wrote: > Hi All, > > We are running informix 7.13 on HP-UNIX 10.01. We are taking a weekly level > 0 > backup of informix and writing backup to EMC clarrion. We noticed every > time > backup will complete at different time, usually it took 14 to 15 hrs to > complete. We have a database of size 36 GB. > > Is that the normal behavior of ontape to backup 36Gb of data in 14 to 15hrs > or > we can do something to complete it faster? Secondly, can anybody help me to > troubleshoot why the backup time varies every time. > > Regards > Deepak > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --0015173fe73874cd9d0482669b06

Deepak, I would have to say that I would expect you have very old hardware if your backup time is that long for a 36GB DB. Using ontape I can backup a 148GB database using a whole system backup (single threaded) in less then 2 hours. What hardware hardware connects you to your disk (DB and Backup)? I'm thinking disk access is your problem. Either you don't have enough bandwidth to get the backup done quickly or you have other large things running at the same time. ie: Local Disk is on 2G Fiber Backup Disk is iscsi across a metro Ethernet I would expect this to run slower then the following Local DB Disk is on 2G Fiber Backup Disk is LTO3 connected local via scsi or 2G Fiber (different controller) Maybe we can help if we had more detail. I would suggest getting someone in to review that hardware and software if you cannot figure it out. Eric B. Rowell On Mon, Mar 22, 2010 at 12:50 PM, DEEPAK JOSHI <djoshih@hotmail.com> wrote: > Hi All, > > We are running informix 7.13 on HP-UNIX 10.01. We are taking a weekly level 0 > backup of informix and writing backup to EMC clarrion. We noticed every time > backup will complete at different time, usually it took 14 to 15 hrs to > complete. We have a database of size 36 GB. > > Is that the normal behavior of ontape to backup 36Gb of data in 14 to 15hrs or > we can do something to complete it faster? Secondly, can anybody help me to > troubleshoot why the backup time varies every time. > > Regards > Deepak > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Eric B. Rowell

Hmm, did you mean 7.31 or 7.13? Don't specifically remember a 7.13, but versions prior to 7.30 also had problems during high update load with slow archives. Also, early 7.xx releases (IB prior to 7.31xD8) experienced the very old page bug which caused archives to run very slowly if the page timestamps wrapped around past 2^31 (or more precisely wrapped so that there were existing pages that have not been modified since the timestamp was last set the such a low number. In this case the engine would have to update every old page on disk that had a timestamp lower than the current wrapped value when it was encountered during the archive. Once it started the updates would go on for many months of archives before there were no longer any pages with a timestamp old enough to cause a problem. It would then reemerge several months later when the timestamps caught up to the next oldest batch of pages. Later releases fixed this performing the timestamp updates separately from the archives. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) IIUG Board of Directors (art@iiug.org) See you at the 2010 IIUG Informix Conference April 25-28, 2010 Overland Park (Kansas City), KS www.iiug.org/conf 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, Mar 22, 2010 at 12:58 PM, Art Kagel <art.kagel@gmail.com> wrote: > Backup times will depend on the IO throughput of the subsystem to which it > is being written and of that of the chunks that are being backed up. Also > if the chunks are on some of the same physical structures that the archive > is being written to, there will be contention for the disk heads. Finally > the configuration of the EMC RAID group that you are writing the archive > file to will affect the back up throughput (ie RAID5 is 40% slower than > RAID10 for sustained sequential writes which is what the archive is doing). > > Differences in archive times day to day can be due to activity on the > server - more updates/inserts/deletes will require that more pre-image data > from the physical log be copied out to the archive after each dbspace's back > up is completed to maintain consistency of the archive. Another cause can > be contention with other applications for IO bandwidth to the EMC array or > even between the machine that is hosting the IDS instance and other machines > on the network that may be sharing the same EMC array. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > IIUG Board of Directors (art@iiug.org) > > See you at the 2010 IIUG Informix Conference > April 25-28, 2010 > Overland Park (Kansas City), KS > www.iiug.org/conf > > 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, Mar 22, 2010 at 12:50 PM, DEEPAK JOSHI <djoshih@hotmail.com>wrote: > >> Hi All, >> >> We are running informix 7.13 on HP-UNIX 10.01. We are taking a weekly >> level 0 >> backup of informix and writing backup to EMC clarrion. We noticed every >> time >> backup will complete at different time, usually it took 14 to 15 hrs to >> complete. We have a database of size 36 GB. >> >> Is that the normal behavior of ontape to backup 36Gb of data in 14 to >> 15hrs or >> we can do something to complete it faster? Secondly, can anybody help me >> to >> troubleshoot why the backup time varies every time. >> >> Regards >> Deepak >> >> >> >> ******************************************************************************* >> Forum Note: Use "Reply" to post a response in the discussion forum. >> >> > --0015173fe7384efb38048266bdd7

DEEPAK JOSHI wrote: > Hi All, > > We are running informix 7.13 on HP-UNIX 10.01. We are taking a weekly level 0 > backup of informix and writing backup to EMC clarrion. We noticed every time > backup will complete at different time, usually it took 14 to 15 hrs to > complete. We have a database of size 36 GB. > > Is that the normal behavior of ontape to backup 36Gb of data in 14 to 15hrs or > we can do something to complete it faster? Secondly, can anybody help me to > troubleshoot why the backup time varies every time. > > Regards > Deepak as you have very often been told in the past year or so, you are running software that's 15 years old, on an operating system that's 15 years old and hardware that's 15 years old (36GB in 14 hours is a bit of a giveaway there). On this specific occasion it's likely that you are running into PTS 101062 (archive performance degradation with too many calls to arc_very_old_page(), or other wording to that effect). Upgrade. -- Ciao, Marco ______________________________________________________________________________ Marco Greco /UK /IBM Standard disclaimers apply! Structured Query Scripting Language http://www.4glworks.com/sqsl.htm 4glworks http://www.4glworks.com Informix on Linux http://www.4glworks.com/ifmxlinux.htm