[SOLVED] Parallel startup for LXC Containers possible?

[Sebastian Schubert]

Hi there,

having >200 LXC Containers on a Server takes some time to bring all up on a Node Reboot, especially when they are started in a serial manner.
Is there a possibility to work with parallel starts to bring up the containers faster?

 

[fabian]

you could write it yourself, either with a script that gets called by a systemd unit, or with a systemd template unit and target. the current onboot start code is not that easy to make parallel, as we have to take configured ordering and startup delays into account. if you don't need any of that, you can simply not set onboot and start them on boot yourself see the pve-guests.service unit for ordering (or simply order your unit after it).

 

[Sebastian Schubert]

We implemented it like this (the pve-guests still takes care of all containers as well, but we dont mind )

Systemd File:

# /etc/systemd/system/startupbooster.service
[Unit]
Description=PVE startup booster
ConditionPathExists=/usr/sbin/pct
RefuseManualStart=true
RefuseManualStop=true
Wants=pvestatd.service
Wants=pveproxy.service
Wants=spiceproxy.service
Wants=pve-firewall.service
Wants=lxc.service
After=pveproxy.service
After=pvestatd.service
After=spiceproxy.service
After=pve-firewall.service
After=lxc.service
After=pve-ha-crm.service pve-ha-lrm.service

[Service]
ExecStartPre=-/usr/share/pve-manager/helpers/pve-startall-delay
ExecStart=/usr/local/bin/startupbooster
Type=oneshot
RemainAfterExit=yes
TimeoutSec=infinity

[Install]
WantedBy=multi-user.target

And the Script itself:

#!/bin/bash
parallel=4
pct list |grep stopped |awk '{print $1}' |  \
while read container; do
        # paralell x 4
        while [ $( jobs -pr | wc -l ) -ge ${parallel} ]
        do      sleep 0.5
        done
        # check if onboot yes/no
        onboot=$(pct config $container|grep onboot|awk '{print $2}')
          if [ "$onboot" == "1" ]; then
           pct start $container &
         sleep 0.5
        fi
done