Search results

Maybe you can try the elevator and the IO scheduling ? https://pve.proxmox.com/wiki/IO_Scheduler + ionice 8 in /etc/vzdump.conf

Hello. From my experience, monitoring software database is the worst case scenario for IOs. Furthermore in your case : small writes (thus read, modify, write), with synchronous random writes, and with synchronous network replication on other nodes ! Maybe you are experiencing IOs starvation on...

Hmmm... And what activity makes those IOs ?

Indeed, nut-scanner does not need nut-snmp.

Hello. All the needed configuration is in /etc. Personally I backup /etc /root /opt to be sure to get any customization I would have made. You can use tar and scp to get the backup and automatize it with Ansible for example. You can use tools like rsnapshots to backup the configuration files...

Hello. I think the SNMP NUT driver is in this separated package and that it is needed. Personally, I let Proxmox host shutdown the CT/VMs in accordance with the priority I defined in Proxmox VM/CT settings. thus I configure NUT only on proxmox hosts, not in VM/CTs.

Nice to read this ️

Sorry, package name was not good (my mistake) : apt install nut-snmp man snmp-ups

Ohhh... Good to know fiona, thanks !

Thank you Fiona. I edited my post. Yest it is the first thing to try, I agree.

So Unspec, maybe you should try to change the scheduler algorithm to CFQ ? nano /etc/default/grub GRUB_CMDLINE_LINUX_DEFAULT="... elevator=bfq" update-grub

Bandwith Bandwidth should apply but maybe IOs will still be high. Why do you think ionice 8 won't do anything ? From the documentation : https://pve.proxmox.com/pve-docs/vzdump.1.html : ionice: <integer> (0 - 8) (default = 7) Set IO priority when using the BFQ scheduler. For snapshot and...

Ok. Indeed, this won't help you when backuping containers.

I think you didn't search too much @greenhornet ... apt install snmp-ups man snmp-ups

Look at "advanced" tab in the backup configuration, and change bandwidth limit (first, try to take 75% of the used bandwidth during backups). But with containers, I fear that IOs will be the problem (plenty of small files)... So maybe try to set the IO workers to 4 ? I never tried this.

Hello, if you can't afford a third server or doubt of CEPH stability with a 3rd server being different... Why wouldn't you put an effort over SQL database clustering ? it seems it can use WSFC, Windows Server Failover Cluster. I didn't dig into it but this could be a solution. At the begining of...

Hello. When I see : snapshot 'vzdump' was not (fully) removed - zfs error: cannot destroy snapshot rpool/data/subvol-116-disk-0@vzdump: dataset is busy It seem to me that the filesystem is overbusy. What do the IO wait graph (MAX) shows during the backup ? Maybe you can try to limit the...

Hello. I use extensively Mikrotik hardware. But I never used MLAG. What you describe seems normal. In documentation (https://help.mikrotik.com/docs/spaces/ROS/pages/67633179/Multi-chassis+Link+Aggregation+Group) : heartbeat (time: 1s..10s | none; Default: 00:00:05) This setting controls how...

Small reminder : do not use ZFS over hardware RAID.

Sorry for you.