Re: Testing backups ...

Dirk, You should test your backups, however your resources may not permit it on the system your concerned with... it's a tough sell to get the disk to restore a large DB. However if you use a storage mgr for multiple systems/databases. Do a test restore on a smaller system, one that you have the resources to allocate the disk for, and validate your backup & restore architecture and methodology in the smaller scale. No that isn't exactly what youv'e been asked to do, but it's a good mitigating controil to ensure your getting solid backups. Rob, Smack your storage guy upside his head with a frying pan, and either get the disk you need, or scrap the HA program and save some dough. It sounds like that admin just isn't looking to do any work that isn't visible to their boss. my 2 cents.... -Chris Rob Vorbroker <rvorbroker@yahoo.com> Sent by: owner-informix-list@iiug.org 08/04/2004 09:00 AM Please respond to rvorbroker@yahoo.com To Dirk Moolman <DirkM@mxgroup.co.za>, informix-list@iiug.org cc Subject Re: Testing backups ... Yes I think you should - preferably on a TEST system of similar configuration. Document your results. I also think we should routinely check HA and DR initiatives but quite often I'm told I live in a dream world. I have a current situation (I'm the new guy so no throwing stones) where the primary and HA database servers have been kept in-step DB resource wise for over a year but perhaps the Unix System parameters, etc are not in sync. Now the storage manager doesn't want to allocate any more space to the HA database server as he states we can't PROVE that it would work. To his point they have NEVER tested whether or not the Secondary server 'could' take over if the Primary went down. I can prove that if we don't add space to the secondary DB when we add space to the primary DB that it won't work... Rob --- Dirk Moolman <DirkM@mxgroup.co.za> wrote: > > > My company wants me to test my backups once a month, > to verify that they still work and can be restored. > > > > How many of you currently do this ? Any good > tips ? > > > > > > We use IDS7.31FD3, on Solaris 9, and the backups are > done using ontape (for Informix), and ufsdumps for > Unix (filesystems) > > > > > > Dirk > > > ________________________________ > > << ella for Spam Control >> has removed 816 Spam > messages and set aside 156 Newsletters for me > You can use it too - and it's FREE! > www.ellaforspam.com > ===== Rob Vorbroker Phone: 513/336-8695 Vorbroker Consulting, Inc. Fax: 513/336-6812 www.vorbroker.com robv@vorbroker.com sending to informix-list sending to informix-list

chris.staubin@reebok.com wrote: > > Dirk, > > You should test your backups, however your resources may not permit it on > the system your concerned with... it's a tough sell to get the disk to > restore a large DB. No longer true. Last prices seen here for 4 TB FIBRE to SATA RAID, 3HE 19" (gives me approx 1.9 TB when configured as RAID 10) was less than 17K USD + VAT. I can read 6500 2KB pages per second (sustained) & write 3350 2KB pages per second, which is not bad for this small amount of $$$. Machine is a 4 CPU / 8 GB domain in a SunFire12K, running Solaris7 and IDS 9.40.UC3W1. google for SATA RAID or - among many other manufacturers - www.infortrend.com (that is what I use atm). Such storage units are available smaller (2 TB) also and with SCSI160 or SCSI320 interfaces as well. Then you pay like 3K USD less. Money to spend is peanuts compared to SANs from EMC or Hitachi or IBM or even HP. BTW, www.storagesearch.com has it all and more .... [ ... rest of valuable text snipped ... ] dic_k -- Richard Kofler SOLID STATE EDV Dienstleistungen GmbH Vienna/Austria/Europe