I know that this is not xfs discussion but last question
With xfs my avg boot up performance is under 30-35s without any vm started.
if you have free time, can you run systemd-analyze on dom0 without any auto start vm?
I know that this is not xfs discussion but last question
With xfs my avg boot up performance is under 30-35s without any vm started.
if you have free time, can you run systemd-analyze on dom0 without any auto start vm?
can you run systemd-analyze on dom0 without any auto start vm?
[user@dom0 ~]$ sudo systemd-analyze
Startup finished in 3.281s (kernel) + 5.351s (initrd) + 8.789s (userspace) = 17.422s
Lenovo ThinkPad T430
i7-3840QM
Samsung V-NAND SSD 860 PRO
time to start a qube = 10 sec
Without (firmware) and (loader) it just take me 13s to boot, and avg 6-8s for vm.
But I surprised even with old machine it almost beat :
I’m interested in benchmarking btrfs next day and sorry for being oot.
avg 6-8s for vm
Well, it depends how I measure.
[user@dom0 ~]$ date && qvm-start qube && date
Fri Oct 15 19:35:43 CDT 2021
Fri Oct 15 19:35:51 CDT 2021
→ 8 sec
[user@dom0 ~]$ date && qvm-run --pass-io qube "echo hello" && date
Fri Oct 15 19:39:04 CDT 2021
hello
Fri Oct 15 19:39:14 CDT 2021
→ 10 sec
How do you measure? My point being: I believe I/O performance is of much higher importance than how many instructions per second your CPU can provide.
Now it kind of is (I split the threads).
Have try benchmarking today, using btrfs increase 2 second in dom0, and 1 sec in vm, haven’t try with large vm storage.