[SOLVED] Wiped removed volumes painfully slow

jlauro

Active Member
Feb 10, 2024
157
48
28
Using PVE 8.4. For LVM iSCSI there is an option for wiped removed volumes. Ideally it would use discard instead, but appears to use /usr/bin/cstream -i /dev/zero which I guess is ok as not all devices support discard (but many do). Unfortunately it also has a -t option that limits it to a ridiculously low speed of 10MB/sec. If it's going to default to something, it show be something a bit more sane, like 100x that. Is there a way to switch it to either use discard instead, or to remove the limit and/or change it to something reasonable?
 
  • Like
Reactions: jlauro
I hope we're talking about the same thing now (saferemove) [0]? The speed can be changed, e.g., 50MB/s.

Code:
pvesm set <lvm-storagename> -saferemove_throughput -52428800


[0] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#pvesm_lvm_config

Yes, that appears to be what I was talking about. Thank you, that worked. I used:
Code:
pvesm set iSAN8V1 --saferemove_throughput 1048576000
(about 1GB/sec)
and it was able to wipe a vm with a 120GB vm in under 2 minutes, so that should be able to wipe a TB in 20 minutes instead of taking over a day.

I should read over some of those admin pages. There might be some other useful settings not exposed in the web interface.
 
  • Like
Reactions: mariol