Hi, need to better understand backups.
i have a single backup setup for all my vms (7kvm, 1 openvz) and got logs like this every morning:
the server is an ibm x3650m2 with 2 hw mirrored local sata disks
disk storage is lvm/scsi over a 1gb eth connection to a qnap ts809u
backup storage is nfs over a 1gb eth connection to a qnap ts809u
the 1gb eth connection is on the same nic
qnap ts809u has 8 1,5TB sata2 disk in sw raid5
proxmox2:~# pveperf
CPU BOGOMIPS: 72537.76
REGEX/SECOND: 788290
HD SIZE: 16.49 GB (/dev/mapper/pve-root)
BUFFERED READS: 151.74 MB/sec
AVERAGE SEEK TIME: 4.21 ms
FSYNCS/SECOND: 1721.37
DNS EXT: 32.65 ms
DNS INT: 0.75 ms (domain omitted)
I know that suspend (0-1 seconds) is invoked when more than 1 disk per vm has to be copied
I can't understand
* why i get those speeds
- are they good, for my setup? i feel they're slow
- shoud them be better?
- how could i improve them? easy/cheap ways?
* why i get different speeds (between 5 and 12) is it about different compression rates? or else?
Thanks to anyone who cares...
Marco
i have a single backup setup for all my vms (7kvm, 1 openvz) and got logs like this every morning:
Code:
|------------------------VMS-------------------------------------------||---------backup setup-------------------||------backup log------------------------|
[FONT=courier new]VMID type os disk n. disk GB disk drv storage mode compress storage status mode suspend MiB/s
102 kvm ubuntu server 2 8+8 IDE lvm/iscsi snapshot yes nfs running snapshot yes 5,10
103 kvm ubuntu server 2 8+8 IDE lvm/iscsi snapshot yes nfs running snapshot yes 5,80
104 kvm ubuntu desktop 1 8 IDE lvm/iscsi snapshot yes nfs running snapshot no 9,18
106 kvm win2003 2 5+15 VIRTIO lvm/iscsi snapshot yes nfs running snapshot yes 10,58
107 kvm ubuntu desktop 2 12+12 IDE lvm/iscsi snapshot yes nfs running snapshot yes 10,64
108 openvz ubuntu 1 2 local local snapshot yes nfs running snapshot no 11,00
109 kvm win2000 1 32 IDE lvm/iscsi snapshot yes nfs running snapshot no 11,42
111 kvm ubuntu server 1 32 IDE lvm/iscsi snapshot yes nfs running snapshot no 7,09[/FONT]
the server is an ibm x3650m2 with 2 hw mirrored local sata disks
disk storage is lvm/scsi over a 1gb eth connection to a qnap ts809u
backup storage is nfs over a 1gb eth connection to a qnap ts809u
the 1gb eth connection is on the same nic
qnap ts809u has 8 1,5TB sata2 disk in sw raid5
proxmox2:~# pveperf
CPU BOGOMIPS: 72537.76
REGEX/SECOND: 788290
HD SIZE: 16.49 GB (/dev/mapper/pve-root)
BUFFERED READS: 151.74 MB/sec
AVERAGE SEEK TIME: 4.21 ms
FSYNCS/SECOND: 1721.37
DNS EXT: 32.65 ms
DNS INT: 0.75 ms (domain omitted)
I know that suspend (0-1 seconds) is invoked when more than 1 disk per vm has to be copied
I can't understand
* why i get those speeds
- are they good, for my setup? i feel they're slow
- shoud them be better?
- how could i improve them? easy/cheap ways?
* why i get different speeds (between 5 and 12) is it about different compression rates? or else?
Thanks to anyone who cares...
Marco
Last edited: