RikSaunderson's questions

I'm using HAProxy to do automatic failover for an LDAP server, but I don't want automatic fail-back.

The scenario is that I have 2 nodes, s1 and s2. I want all traffic going to s1 unless it fails, when it fails, I want all traffic going to s2. When s1 becomes healthy again, I don't want to automatically route traffic back to it because some entries will have been created on s2 when it was live. I want to ensure that all data has been synchronised before I make s1 active again.

I'm configuring HAProxy like so:

backend bk_app
    stick-table type ip size 1 nopurge
    stick on dst
    option httpchk
    server s1 10.37.226.118:80 check
    server s2 10.37.226.113:80 check backup

This appears to work, connections continue to go to s2 even after s1 is healthy again.

My question is - how do I clear the stick-table from the command line, in other words how do I redirect traffic back to s1?

I have seen some advice which says that if I force s2 to fail, HAProxy will redirect traffic back to s1. This seems to work, but when s2 becomes healthy again, HAProxy directs traffic back to s2 again, in other words the stick-table is still in effect.

I have seen some advice which says that I just need to restart HAProxy, which works perfectly, but this does seem like the nuclear option! Is there some sort of command line option to clear the stick-table?

I'm trying to set up a high availability cluster in AWS for an application that I'm running. The application is on an EC2 instance which sits in front of an Oracle RDS. If I use traditional load balancing and have two instances running at the same time, the database becomes corrupt because the two instances will be making changes to the database but will not be aware of each other, so I need to do active/passive load balancing.

The problem that I have is that if the passive instance is running the application, even with no traffic coming into it, it nevertheless makes changes to the database around 0.1% of the time. This means that I need to ensure that the application service is not running on the backup node when the primary node is healthy.

My ideal scenario would be to have something that runs active/passive failover which can do the following:

1. Health-check the primary node
2. If the primary node is healthy, forward traffic to the primary node
3. If the primary node is not healthy, run a script which starts up the service and then forward traffic to the secondary

I've been looking into HAProxy, but I haven't seen a way to get HAProxy to run an arbitrary script on the backup server in the event of a failure in the primary node.

I've seen some discussion on using keepalived. Is what I'm suggesting possible in keepalived? Is there something else that can do this?

I have a site.pp file which looks like this:

Package {allow_virtual => false,}

stage { 'pre': before => Stage['main'] }
stage { 'post': require => Stage['main'] }

resources { 'firewall': purge => true }

node default {
  hiera_include('classes')
}

I’d like to be able to set the firewall purge line based on the value of a hiera parameter. Is this possible?

I'm using VSphere to spin up a series of virtual machines with CentOS 6.5 as the OS.

I'd like to be able to tell, in a bash script, once the OS installation has completed. I've tried using ping, but that returns true while Anaconda is still running.

Is there a way using either bash or the VSphere API to determine whether the OS installation is complete?

I’m having some trouble setting up bonded interfaces in a CentOS 6.5 kickstart script. I have 8 physical network interfaces, which want to be configured in pairs as 4 bonded network interfaces.

My network settings are as below:

network --device=bond0 --noipv6 --bootproto=static --onboot=yes --ip=172.123.1.22 --netmask=255.255.255.0 --bondslaves=eth0,eth1 --bondopts=mode=active-backup,primary=eth0,miimon=80,updelay=60000

network --device=bond1 --noipv6 --bootproto=static --onboot=yes --ip=172.123.2.22 --netmask=255.255.255.0 --bondslaves=eth2,eth3 --bondopts=mode=active-backup,primary=eth2,miimon=80,updelay=60000

network --device=bond2 --noipv6 --bootproto=static --onboot=yes --ip=172.123.3.22 --netmask=255.255.255.0 --bondslaves=eth4,eth5 --bondopts=mode=active-backup,primary=eth4,miimon=80,updelay=60000

network --device=bond3 --noipv6 --bootproto=static --onboot=yes --ip=172.123.4.22 --netmask=255.255.255.0 --bondslaves=eth6,eth7 --bondopts=mode=active-backup,primary=eth6,miimon=80,updelay=60000

network --device=eth0 --noipv6 --nodns --bootproto=static --onboot=yes

network --device=eth1 --noipv6 --nodns --bootproto=static --onboot=yes

network --device=eth2 --noipv6 --nodns --bootproto=static --onboot=yes

network --device=eth3 --noipv6 --nodns --bootproto=static --onboot=yes

network --device=eth4 --noipv6 --nodns --bootproto=static --onboot=yes

network --device=eth5 --noipv6 --nodns --bootproto=static --onboot=yes

network --device=eth6 --noipv6 --nodns --bootproto=static --onboot=yes

network --device=eth7 --noipv6 --nodns --bootproto=static --onboot=yes

My problem is this: bond0 is set up correctly, but bond1, bond2 and bond3 are not set up.

I have seen a number of messageboard posts on similar topics, and the resolution always seems to be to put in a custom bash script for the purpose of setting up the bonded interfaces. However, I don’t think I believe that this is the only way to do it. Why would there be a mechanism in the kickstart file for setting up bonded interfaces if it doesn’t work properly?

I’m attempting to use the VMWare command line interface to mount a datastore ISO file in the CD/DVD drive of a virtual machine. An entry elsewhere on serverfault tells me that ‘vmware-cmd blabla.vmx setconfig ide0:0.fileName cdrom.iso’ ought to do what I need, but my installation of the command line tool does not have the setconfig command.

Further research shows that the setconfig command has been removed from the command line tools, and there are even some suggestions that vmware-cmd is deprecated.

I'm developing an automated deployment/configuration system using puppet, and my aim is to test that system on virtualised infrastructure as a nightly build on Jenkins.

Our Jenkins server is a linux server, so we can't use the VMWare Power CLI (which is Windows only).

Has anyone tried to mount an iso on an ESXI virtual machine from the Linux command line in the recent past? If so, how did you do it?

We have a server which has trouble booting normally. The server is an HP Proliant DL380G6, running SuSe Enterprise Linux 10 Service Pack 2. When we try to boot it, it goes into some sort of disk check. The disk check gets stuck half way through and never completes, meaning that we never get through to the actual OS. Rebooting causes the same issue.

When we try to boot into failsafe mode, the disk check issue goes away. Obviously this is not a long term fix, but for the short term, this does give us a functioning server. However, there are times when we need to perform an unattended reboot of the box (for example if the power goes out).

Is there a way of configuring our server so that it boots into failsafe mode by default?

Thanks, Rik

I have set up a linux (CentOS 5) box with a samba share on a large network. The network has 2 subnets, one on 192.168/16 and one on 172.18/16. The server lives on the 172.18/16 subnet.

I can ping the server from client Windows XP workstations on both subnets.

When I open the list of workgroup computers, I can see my new server in the list of Workgroup machines from Windows XP client workstations on both subnets.

However, I can access the share from workstations on the same subnet, but when I try to access it from workstations on the other subnet I get a message saying either:

o Windows cannot find the specified resource, or
o Access to this network resource has been disallowed.

I have had a look at the smb.conf file, and there is nothing obvious that would be causing a problem.

What am I doing wrong? Thanks, Rik