Results 1 to 9 of 9

Thread: OpenVZ has problems with FC10 (sudo and "bad address")

  1. #1
    Join Date
    Apr 2009
    Posts
    19

    Exclamation OpenVZ has problems with FC10 (sudo and "bad address")

    I've ran into a couple of problems with running the FC10 template on the current OpenVZ.

    My VPS machine..

    [root@servername ~]# uname -a
    Linux servername.mydomain.com 2.6.18-ovz028stab053.5-smp #1 SMP Wed Mar 26 12:01:19 PDT 2008 i686 i686 i386 GNU/Linux


    THE TWO PROBLEMS...
    I am unable to post URLs, so I've used tinyurl [dot] com. Please add the character code to tinyurl [dot] com [slash] [character-codes] to view the URL.


    1) Setting up a user to perform a "sudo" command under another user seems to be broken in the current kernel running my instance.

    [root@servername ~]# sudo -u ftp ls /
    sudo: Error dropping capabilities, aborting : Operation not permitted

    This is documented in the forum posts that I've made on "form [dot] openvz [dot] org"

    OpenVZ Forum: Support => SUDO inside VPS don't work
    tinyurl [dot] com [slash] d847gl


    2) The "touch" and "cp -p" commands are returning an error message when utilized.

    [root@servername ~]# touch test.txt
    touch: setting times of `test.txt': Bad address
    Last edited by boulderdrop; 04-14-2009 at 04:48 PM.

  2. #2
    Join Date
    Dec 2007
    Posts
    1,141

    Default

    Quote Originally Posted by boulderdrop View Post
    I've ran into a couple of problems with running the FC10 template on the current OpenVZ...
    I did not see any tickets in our system regarding these OS template issues, however, I have created a ticket for immediate consideration from our template developers.


    Quote Originally Posted by boulderdrop View Post
    I am unable to post URLs...
    My apologies for the inconvenience - we have seen a sharp reduction in spam after implementing the linking restriction (so it is here to stay), however, we do plan to implement a spam prevention workaround for existing VPSLink customers to eliminate the link restriction.

  3. #3
    Join Date
    Apr 2009
    Posts
    19

    Default

    Thanks. Moved my responses to the ticket which was created.

  4. #4
    Join Date
    Apr 2009
    Posts
    1

    Default Same Issue

    I'm having the same issue described above. I've traced it down to an update of coreutils, which is expecting a newer kernel version.

    Is there a resolution to this issue?

    Edit: Running Gentoo VPS, if that helps.

  5. #5
    Join Date
    Apr 2009
    Posts
    19

    Default

    I'm not really sure how much help you can expect to get on this.

    I've been sending in Ticket Responses for this problem since Apr 14th.

    Today, I saw that they responded with this:
    "I still think he's crazy, but whatever. sa ticket already made. stalling."

    Apparently they didn't realize that customers can see ticket-replies.

    I've made posts to the forums at OpenVPS and it looks like it could be a configuration issue with the hosting-kernel; but who knows, the people at VPSLink don't seem very interested in helping. Sigh.

  6. #6
    Join Date
    Jul 2007
    Location
    127.0.0.1
    Posts
    392

    Smile

    Quote Originally Posted by boulderdrop View Post
    Today, I saw that they responded with this:
    "I still think he's crazy, but whatever. sa ticket already made. stalling."
    Despite saying "FC10" in your initial post, I don't think you're crazy - Fedora Core was discontinued after release 6 (FC6).

    Quote Originally Posted by boulderdrop View Post
    I've made posts to the forums at OpenVPS and it looks like it could be a configuration issue with the hosting-kernel; but who knows, the people at VPSLink don't seem very interested in helping. Sigh.
    After taking a look at the thread it's pretty obvious there's an issue, even with the older F8 release. I know this doesn't address the overall problem, but perhaps a workaround is in order until a known solution is determined?

    Consider one of the many options VPSlink provides:
    Quote Originally Posted by vpslink
    Debian
    * Debian 4 "Etch"
    * Debian 4 w/Ruby on Rails (Apache)
    * Debian 4 w/Ruby on Rails (nginx)
    * Debian 4 w/LAMP
    * Debian 5 "Lenny"

    Ubuntu
    * Ubuntu 6.06 LTS (Dapper Drake)
    * Ubuntu 8.04 LTS (Hardy Heron)
    * Ubuntu 8.04 LTS (Hardy Heron) w/LAMP
    * Ubuntu 8.10 (Intrepid Ibex)

    CentOS
    * Cent*S 4
    * Cent*S 4 w/LAMP
    * Cent*S 5
    * Cent*S 5 w/LAMP
    * Cent*S 5 w/HostInABox/LXAdmin

    Fedora
    * Fedora 9 "Sulphur"
    * Fedora 9 w/LAMP

    Slackware
    * Slackware 11.0
    * Slackware 12.0
    * Slackware 12.1

    OpenSUSE
    * OpenSUSE 10.3
    * OpenSUSE 11.0

    Arch Linux
    * Arch Linux 0.8

  7. #7
    Join Date
    Dec 2007
    Posts
    1,141

    Default

    Our system administrators have confirmed that there is a known issue affecting the exit code for the listed commands owing to a kernel mismatch between the kernel version running on our hardware nodes (presently the CentOS 5 stable kernel) and the expected kernel for newer Arch, Fedora, and Gentoo releases.

    A resolution to this issue is pending further investigation on the part of our system administrators: until the stable CentOS kernel is updated, a patch or kernel update must be implemented to correct this behavior.

    Our system administrators report that - to date - patches which have been tested have failed to correct this issue without introducing new bugs.

    We are presently preparing a realistic upgrade path for all OpenVZ hardware nodes to address this issue.

  8. #8
    Join Date
    Jul 2007
    Location
    127.0.0.1
    Posts
    392

    Unhappy

    Quote Originally Posted by DanL@VPSLink View Post
    We are presently preparing a realistic upgrade path for all OpenVZ hardware nodes to address this issue.
    Thanks for shining some light on the issue, Dan. Just a quick question..

    You mentioned "all OpenVZ hardware nodes" despite the fact not all templates display or share the issue initially mentioned by boulderdrop. Does this mean customers using non affected templates will eventually experience a reboot?

    I apologize in advance if I'm asking the obvious, just wanting to make sure.

  9. #9
    Join Date
    Feb 2006
    Posts
    773

    Default

    We will likely be updating *all* of our hardware nodes. So at some point all OpenVZ customers will experience a reboot. Sorry :/

Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •