Hi,
we use CIFS Protocol for Backup, but the check if the Storage is online produces high load on our Samba file servers.
The samba fileserver is bound to our Active Directory, in the log you can see, that the online checker tries an login every 10 seconds:
[2019/01/28 13:46:13.847175, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-bck-cfw01a]"
[2019/01/28 13:46:13.847368, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-bck-kp]"
[2019/01/28 13:46:13.847550, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-backup]"
[2019/01/28 13:46:13.847735, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-iso]"
[2019/01/28 13:46:13.950122, 2] ../source3/auth/auth.c:305(auth_check_ntlm_password)
check_ntlm_password: authentication for user [pve-backup] -> [pve-backup] -> [AD\pve-backup] succeeded
as we have an large number of Proxmox Servers, which all try to check the load on the File servers goes up:
[root@gar-sv-pool03b samba]# top
top - 13:54:00 up 81 days, 4:29, 1 user, load average: 4.41, 3.78, 3.26
Tasks: 50 total, 2 running, 48 sleeping, 0 stopped, 0 zombie
%Cpu(s): 7.4 us, 3.9 sy, 0.0 ni, 86.6 id, 1.8 wa, 0.0 hi, 0.3 si, 0.0 st
KiB Mem : 49278476 total, 11101584 free, 16559780 used, 21617112 buff/cache
KiB Swap: 15572988 total, 15473916 free, 99072 used. 15107436 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
17221 root 20 0 4113036 2.049g 2.047g R 63.7 4.4 0:03.15 smbd
17225 root 20 0 4106644 923324 920844 S 57.3 1.9 0:01.72 smbd
17224 root 20 0 4106644 921372 918892 S 56.7 1.9 0:01.70 smbd
17228 root 20 0 2918128 5876 3404 S 22.7 0.0 0:00.68 smbd
8782 pve-bac+ 20 0 4120428 1.611g 1.608g S 1.3 3.4 172:30.98 smbd
8764 root 20 0 2913048 2.326g 2.324g S 0.3 4.9 308:14.99 smbd
It is no option to switch to NFS for the backup, as this kicks the proxmox HA Nodes.
we use CIFS Protocol for Backup, but the check if the Storage is online produces high load on our Samba file servers.
The samba fileserver is bound to our Active Directory, in the log you can see, that the online checker tries an login every 10 seconds:
[2019/01/28 13:46:13.847175, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-bck-cfw01a]"
[2019/01/28 13:46:13.847368, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-bck-kp]"
[2019/01/28 13:46:13.847550, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-backup]"
[2019/01/28 13:46:13.847735, 2] ../source3/param/loadparm.c:2769(lp_do_section)
Processing section "[pool03b_pve-iso]"
[2019/01/28 13:46:13.950122, 2] ../source3/auth/auth.c:305(auth_check_ntlm_password)
check_ntlm_password: authentication for user [pve-backup] -> [pve-backup] -> [AD\pve-backup] succeeded
as we have an large number of Proxmox Servers, which all try to check the load on the File servers goes up:
[root@gar-sv-pool03b samba]# top
top - 13:54:00 up 81 days, 4:29, 1 user, load average: 4.41, 3.78, 3.26
Tasks: 50 total, 2 running, 48 sleeping, 0 stopped, 0 zombie
%Cpu(s): 7.4 us, 3.9 sy, 0.0 ni, 86.6 id, 1.8 wa, 0.0 hi, 0.3 si, 0.0 st
KiB Mem : 49278476 total, 11101584 free, 16559780 used, 21617112 buff/cache
KiB Swap: 15572988 total, 15473916 free, 99072 used. 15107436 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
17221 root 20 0 4113036 2.049g 2.047g R 63.7 4.4 0:03.15 smbd
17225 root 20 0 4106644 923324 920844 S 57.3 1.9 0:01.72 smbd
17224 root 20 0 4106644 921372 918892 S 56.7 1.9 0:01.70 smbd
17228 root 20 0 2918128 5876 3404 S 22.7 0.0 0:00.68 smbd
8782 pve-bac+ 20 0 4120428 1.611g 1.608g S 1.3 3.4 172:30.98 smbd
8764 root 20 0 2913048 2.326g 2.324g S 0.3 4.9 308:14.99 smbd
It is no option to switch to NFS for the backup, as this kicks the proxmox HA Nodes.