LXC - I-MSCP 1.4.x - Debian Wheezy to Jessie

  • Hello,

    I am running Debian Wheezy at the moment and planning to upgrade to Jessie because support was dropped in 1.4.x series. I have LXC environment. I was wondering if i-mscp 1.4.x requires systemd or does it work fine with Jessie using sysvinit?

  • @c64wolf

    i-MSCP has providers for the following init systems, whatever the distribution in use:

    • sysvinit
    • upstart
    • systemd

    So, yes, it will "normaly" work just fine with Debian Jessie with sysvinit as long as there are "fallback" sysvinit scripts provided for all services (which must be the case for Jessie).

    Why not Systemd with Jessie (even on LXC) ?


  • I remember reading from Debian errata that dist-upgrading Wheezy LXC container won't work because it pulls in systemd and breaks things, but I guess it is no longer an issue (looked up the issue and found only this https://www.debian.org/release…n.html#lxc-upgrade-issues). The debian release notes state that I would need to add following to LXC container config for systemd to work in container after upgrade:

    1. lxc.autodev = 1
    2. lxc.kmsg = 0

    I guess this is true for Proxmox host too as it is running systemd and is based on Jessie.

  • I managed to upgrade the VPS to Jessie and to 1.4.3 version of i-mscp.

    I am having issues with email service. Currently it seems that I am not receiving any emails from outside and mail.log is full of this:
    May 19 17:43:44 web postfix/pipe[17219]: F072CC802EF: to=<ccc@example.com>, relay=maildrop, delay=101193, delays=101193/0.01/0/0.25, dsn=4.3.0, status=deferred (Command died with signal 11: "maildrop")

    and this
    May 19 17:42:39 web imapd-ssl: couriertls: /etc/imscp/imscp_services.pem: error:0906D06C:PEM routines:PEM_read_bio:no start line

    Any help? I am using postgrey plugin. As a temporary solution I will disable it now and reboot server.

  • More information, in addition I am getting apparmor errors in /var/log/messages inside the container

    1. May 19 16:09:01 web kernel: [335147.880658] audit: type=1400 audit(1495199341.753:872): apparmor="DENIED" operation="file_lock" profile="lxc-container-imscp" pid=26284 comm="(ionclean)" family="unix" sock_type="dgram" protocol=0 addr=none
    2. May 19 16:18:32 web kernel: [335718.572511] maildrop[28079]: segfault at 0 ip (null) sp 00000000ffee6ccc error 14 in maildrop[8048000+35000]
    3. May 19 16:18:44 web kernel: [335730.707563] maildrop[28103]: segfault at 0 ip (null) sp 00000000ffabd07c error 14 in maildrop[8048000+35000]
    4. May 19 16:23:44 web kernel: [336030.869634] maildrop[29013]: segfault at 0 ip (null) sp 00000000fffa851c error 14 in maildrop[8048000+35000]
    5. May 19 16:28:44 web kernel: [336330.427211] maildrop[29946]: segfault at 0 ip (null) sp 00000000ffefca1c error 14 in maildrop[8048000+35000]
    6. May 19 16:33:44 web kernel: [336630.799952] maildrop[30916]: segfault at 0 ip (null) sp 00000000ff862b4c error 14 in maildrop[8048000+35000]
    7. May 19 16:33:44 web kernel: [336630.960986] maildrop[30920]: segfault at 0 ip (null) sp 00000000ffd3d48c error 14 in maildrop[8048000+35000]
  • host node /etc/pve/CTID.conf configuration for the VPS

    1. arch: i386cpulimit: 3cpuunits: 1024hostname: web.example.commemory: 2560net0: net0: name=eth0,bridge=vmbr0,gw=xx.xx.xx.xx,hwaddr=02:xx:xx:xx:xx:xx,ip=xx.xx.xx.xx/32,type=vethostype: debianrootfs: local:101/vm-101-disk-1.raw,size=500Gswap: 2560lxc.cap.drop:lxc.aa_profile: lxc-container-imscp

    lxc-container-imscp apparmor config

    1. profile lxc-container-imscp flags=(attach_disconnected,mediate_deleted) {
    2. #include <abstractions/lxc/container-base>
    3. mount,
    4. remount,
    5. umount,
    6. }
  • Disabling postgrey did not help. This VPS worked fine when it was still Wheezy and i-mscp 1.3.x series. Considering rolling back to backup before the upgrade. :/

  • Restoring the container from backup made everything work again (Wheezy, I-MSCP 1.3.16), so the issue resides somewhere in the upgrade process. Either Jessie works differently from AppArmor point of view or there is a configuration issue somewhere. During the upgrade process I did have to purge old apache and php packages to make the install work. Other than that the upgrade was standard (stop imscp_panel and imscp_daemon, change apt sources list entries to jessie, apt-get update && apt-get upgrade && apt-get dist-upgrade, apt-get install -f, apt-get autoremove, and upgrade of imscp by standard procedure).

  • @c64wolf

    You are using proxmox for managing your lxc containers? If yes, which version?


  • @c64wolf

    You are using proxmox for managing your lxc containers? If yes, which version?

    Thank you for fast reply! Host node /etc/debian_version says 8.8 and Proxmox VE version is 4.4-13/7ea56165. Kernel version is 4.4.59-1-pve #1 SMP PVE 4.4.59-87 (Tue, 25 Apr 2017 09:01:58 +0200) x86_64 GNU/Linux.