Clean install - Proxmox 4.4 High I/O

Mislav

Active Member
Jan 15, 2019
4
0
41
35
Hi everyone. I've noticed that my iotop is showing the following:

TID PRIO USER DISK READ DISK WRITE SWAPIN IO> COMMAND
7252 be/3 root 0.00 B/s 3.83 K/s 0.00 % 99.99 % [jbd2/loop0-8]
249 be/3 root 0.00 B/s 0.00 B/s 0.00 % 99.99 % [jbd2/md2-8]

Doing some debugging with
echo 1 > /sys/kernel/debug/tracing/events/ext4/ext4_sync_file_enter/enable

and checking file
/sys/kernel/debug/tracing/trace_pipe

I've noticed that every send there is sync for mysql and loop0 which is probably single LXC container - currently in RAW format, created through proxmox (version 4,4 - running under Debian 8).

Whole VM is terribly slow and disk i/o is causing problems. What options do I have and how to solve this? I've read on forums that some people switched from raw to chroot and that helped?
 
Just a side question, how is this working under Debian 9 - same raw format or you can choose? Is it still that slow?
 
Storage for LXC containers?
That depends on what you use as storage-plugin, e.g.:
ZFSpool - a dataset for each mountpoint (you can see the directories on the host)
* lvm(thin) - one LV (blockdevice) per pool - which gets formatted with ext4 and then mounted
* ceph - one krbd (blockdevice) - same as above
* directory (seems the one you used above) - raw image as above

see https://pve.proxmox.com/pve-docs/pve-admin-guide.html#pct_container_storage
 
Cool. I like the old way of promox 2 and 3 with openvz containers - that was working fine and never and disk I/O problems. I will check the guide, thanks!