Jira (FACT-2479) Uptime facts on LXC get the uptime from the host OS and not the container.
Title: Message Title
Atanas Stoyanov commented on FACT-2479
Re: Uptime facts on LXC get the uptime from the host OS and not the container.
Hi Florin, I spawned a new test LXC and the facter --version command returns: 3.14.9 (commit 9115836a933a9191729305201b32ed46713ecf1b) I upgraded the facter with "/opt/puppetlabs/puppet/bin/gem install facter" to facter-4.0.33 and the issue seem to have disapeared for the commandline, however when checking the facts through foreman or via the puppet facts command the values are still wrong:
root@LXC # puppet facts |grep uptime
"ia_uptime_seconds": "79",
"system_uptime": {
"uptime": "22 days"
"uptime": "22 days",
"uptime_days": 22,
"uptime_hours": 549,
"uptime_seconds": 1978892,
root@LXC # facter uptime
0:01 hours
I thought I may have another facter binary but didn't find such:
Jira (FACT-2479) Uptime facts on LXC get the uptime from the host OS and not the container.
Title: Message Title Atanas Stoyanov commented on FACT-2479 Re: Uptime facts on LXC get the uptime from the host OS and not the container. Hi, Just a small clarification, we are not running LXD but pure LXC. root@serv # facter uptime 1:00 hours root@serv # # uptime 13:47:36 up 1:00, 4 users, load average: 0.12, 0.09, 0.16 root@serv # lxc-attach -n ubuntu -- facter uptime 1:00 hours root@serv # lxc-attach -n ubuntu -- uptime 13:47:43 up 28 min, 0 users, load average: 0.10, 0.08, 0.16 root@serv # lxc-attach -n ubuntu -- cat /proc/uptime 1694.0
Jira (FACT-2479) Uptime facts on LXC get the uptime from the host OS and not the container.
Title: Message Title Atanas Stoyanov created an issue Facter / FACT-2479 Uptime facts on LXC get the uptime from the host OS and not the container. Issue Type: Bug Assignee: Unassigned Created: 2020/03/20 1:39 AM Environment: OS: Ubuntu Bionic 18.04 Server: Puppetserver 6.7.1-1bionic Agent: puppet-agent 6.10.1-1bionic LXC: lxc 3.0.3-0ubuntu1~18.04.1 Priority: Normal Reporter: Atanas Stoyanov Hello everyone, On LXC the core facts in: system_uptime -> hours system_uptime -> seconds system_uptime -> days system_uptime -> uptime uptime_hours uptime_seconds uptime_days uptime display the uptime of the host OS and not the LXC itself. I am not sure where the uptime is taken and calculated from but a good source of truth would be the /proc/uptime, which displays the proper value inside the LXC. Add Comment
Jira (PUP-10141) Puppetserver 3-4 times more overloaded after upgrade from 5.3 to 6.7.1
Title: Message Title Atanas Stoyanov created an issue Puppet / PUP-10141 Puppetserver 3-4 times more overloaded after upgrade from 5.3 to 6.7.1 Issue Type: Bug Affects Versions: PUP 6.10.1 Assignee: Unassigned Attachments: Screenshot 2019-11-15 at 09.32.35.png, Screenshot 2019-11-15 at 09.32.40.png, Screenshot 2019-11-15 at 09.32.49.png, Screenshot 2019-11-15 at 10.17.18.png Components: Compiler Created: 2019/11/15 1:18 AM Priority: Normal Reporter: Atanas Stoyanov Puppet Version: 6.10.1 Puppet Server Version: 6.7.1 OS Name/Version: Ubuntu Bionic We have a puppet environment of 3 masters and ~250 clients. One of the masters is also serving CA. We have a HA setup with srv records. When we push our puppet code to our internal git, it gets pushed to the puppet masters and there we have different environments depending on the branch. Once we upgraded from puppetserver 5.3 to 6.7.1 the servers started to eat up all the server memory and occasionally die, thus we had to add more ram and CPUs to the VMs, where they are running. The CPU usage of the 3 masters also spiked a lot (see graphs). Also on the clients we noticed a signifficant slowdown in the puppet agent -t, which we assume is caused by the slow catalog compilation on the masters. The client's CPU usage, Memory usage and other stats except for the puppet run time seem unchanged. puppetserver.log does not contain any errors. On the attached images can be clearly seen when the upgrade was performed on 12.11.2019 ~ 11:00 Desired Behavior: Make puppetserver run smooth again
