Deb, Kees, Yes, I'd done something similar, we have zfs mounts for each user, and then globbed them together by first letter of the username.
Brian Extract from disklist ZFS mounts for samba shares was by specific share, I'd written a script to find the shares that currently existed and updated the disklist daily with the current list, automatically so shares could be created and not missed, we seldom took any offline. I did not bother with spindles for this. Got great plots from # amplot which was critical in pinpointing bottlenecks. # extracted from home directory list, new home directories caught automatically by the glob. finsen /export/home-Y /export/home { user-tar2 include "./[y]*" } finsen /export/home-Z /export/home { user-tar2 include "./[z]*" } finsen /export2/home-AZ /export2/home { user-tar2 include "./[A-Z]*" } List supporting samba shares, updated by daily script. finsen /export2/samba/bdinst zfs-snapshot2 finsen /export2/samba/bdlshare zfs-snapshot2 finsen /export2/samba/bladder zfs-snapshot2 -----Original Message----- From: owner-amanda-us...@amanda.org <owner-amanda-us...@amanda.org> On Behalf Of Debra S Baddorf Sent: Tuesday, September 21, 2021 1:38 PM To: Kees Meijs | Nefos <k...@nefos.nl> Cc: Debra S Baddorf <badd...@fnal.gov>; amanda-users@amanda.org Subject: Re: CPU eating tar process ATTENTION: This email came from an external source. Do not open attachments or click on links from unknown senders or unexpected emails. Have you experimented with dividing those target disks into smaller pieces, using tar? So that amanda isn’t doing level 0 on all parts on the same day? I’ve divided some disks as far as a*, b*, c*, …. z*, Other (to catch caps or numbers or future additions). I’ve found that each piece must have SOME content, or tar fails. So Other always contains some small portion, and non-existent letters are skipped and are caught by Other if they’re created later. It’s a pain for restoring a whole disk, but it helps backups. Deb Baddorf Fermilab > On Sep 21, 2021, at 8:55 AM, Kees Meijs | Nefos <k...@nefos.nl> wrote: > > Hi list, > > We've got some backup targets with lots (and lots, and then some) of files. > There's so much of them that making back-ups is becoming a problem. > > During the backup process, tar(1) is eating up a CPU core. There's no or > hardly no I/O wait to be seen. Very likely tar is single threaded so there's > that. The additional gzip(1) process is doing zero to nothing. > > Any thoughts on speeding this up? Maybe an alternative for GNU tar, or...? > > Thanks all! > > Cheers, > Kees > > -- > https://protect2.fireeye.com/v1/url?k=65a54a9a-3a3e73f8-65a7b3af-000babd9069e-d60d8dd67d1ef6d1&q=1&e=de036afa-6a0d-4723-8c1b-1a3c292b5206&u=https%3A%2F%2Fnefos.nl%2Fcontact > > Nefos IT bv > Ambachtsweg 25 (industrienummer 4217) > 5627 BZ Eindhoven > Nederland > > KvK 66494931 > > Bereikbaar op maandag, dinsdag, donderdag en vrijdag tussen 09:00u en 17:00u.