Peter, this is really not something that can be easily answered. It's going to be different from site to site.
Basically, you need to do whatever backups allow you to get back to a point in time that your site can live with. There are sites that take weekly backups on Saturday/Sunday, thinking that if they had a disaster, that they are fine with falling back a week. There are sites that do that and take daily incremental backups with HSM or FDR/ABR, thinking that they are okay with falling back a day. Then there are those that need to have redundant storage arrays, so that if something happens that they only have to fall back a few seconds. Some have redundant virtual tape or rsynced tape because they can't afford to lose any tapes. Then there are the countless sites in between the above that have varying recovery goals. Some can lose some parts of the system back a couple weeks, and then there are other parts that even losing an hour would be catastrophic. You have to decide what your specific goals are at your site. If you have some stuff that can be backed up weekly or monthly (the MVS res and dlibs come to mind), then backing them up daily is a waste of resources. You might have some files that need to be backed up as soon as they are written, and some that can go straight to delete. You have to work with your site and decide what is critical, and what isn't and set up your recovery scheme so that you can meet whatever those goals are. This is one of those things that it's much easier to work backwards from the goal to develop what you need to do in order to meet those goals. There are a lot of really cool things that you can do to meet your goals, but you have to have them first. So, go back and figure out what "HAS TO BE" to continue being able to get work done, and first, make sure that you can do that much. Then figure out what "WOULD BE NICE" to also have, and make a plan for that. As you continue to meet your goals, you will at that same time be building your disaster recovery plan. How long you keep stuff is also going to be part of that plan. If you have data that you need to be able to recover to some older "point in time" so that you can rebuild something, that needs to be part of your plan and it will govern how long you need to keep data, as well as "WHERE" you keep the data. Having incremental backups that you run 24x7 are going to be worthless if you can't use them to recover anything. SO, in this case you need a plan for your site, and no one can really help you do that part, because you have to decide (with the people at your site) what is important. On Wed, 15 Oct 2025 07:57:07 +0400, Peter <[email protected]> wrote: >Hello > >Just trying to understand some of your experience about physical tape >backup up. > >I know almost most of the shops are tapeless. >But when you had physical tape , > >1 ) What are the files you backed up to 3590 ? >2 ) what was the retention period you followed for database , system >volumes and user datasets ? >3 ) Generally to recover an entire lpar from a physical tape a volume level >backup for an entire lpar would suffice ? > > >Any information on the above would help me to research further. > >Regards >Peter > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, >send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
