> On Nov 2, 2016, at 5:16 PM, Stefan G. Weichinger <[email protected]> wrote: > > Am 2016-11-02 um 22:21 schrieb Debra S Baddorf: >> Good to know you’ve done it before. >> >> It takes a looooong time for government to buy anything. A month is >> actually an awfully >> optimistic estimate, now that I think about it! >> >> But I’m told this temporary node has 20+ 2TB disks, and yes I’ll just do >> one full and then incrementals. >> Probably of the user-data areas only, not the OS’s on all 9 or more clients. >> I’ll just keep adding more spool areas as the first setup gets full. My >> current backups >> of these clients (different software; now expired!) is taking 3-4 LTO4 >> tapes for a level 0 backup. > > I would suggest to adjust the "reserve" parameter. > > And maybe the cronjobs: "amdump --no-taper" ... to keep the noise low ;-)
I usually set the reserves to 02% That’s a great idea that I hadn’t yet thought of for the amdump command. I use that when manually re-running a dump of a single node — but hadn’t thought of it for THIS instance. Thanks- Deb
