I'm looking for a way to filter a packet capture in wireshark for instances where our server responds with "Refused" to a recursive DNS query.
dns.resp.type== doesn't seem to offer anything that I see as a match to my request, do I need to look somewhere other than under dns.resp altogether?
I've been trying to figure out how to convert
dnf module reset php
dnf module install php:remi-7.4
to a stanza in a puppet module for several hours without any success.
Has anyone figured out how to do that? The doco on the puppet website is somewhat lacking, shall we say.
I've been facing an interesting issue for a few weeks now. Part of our fleet is only reachable via a jump-host.
In the past I had no issues w/ logging into the jump-host from my workstation using ssh -A and maintain a session for extended periods of time [yes, I know that's bad practice, not here for a lecture on a foul habit =}], and connect on from there to other machines w/o password entry as expected.
Now every so often going to my uxterm with the active session to the jump host gives me a
Permission denied (publickey)
If I disconnect, and w/o restarting the agent connect to the jump host again, all works as before.
How do I diagnose what's going on there?
Extra information: my workstation is Ubuntu 16.04, the jump host is 14.04, recently upgraded from 12.04.
After lots of trial & error I've finally given up on the "SATA RAID" the R320 provides, and resorted to using AHCI mode.
Partitioning/Provisioning work fine, but on the next boot the machine claims to have no OS.
Booting it in rescue mode of an ISO (vFlash) I can mount the disks, chroot to the installed OS, re-install GRUB w/o errors, but it still won't boot afterwards.
Running a complete install of the ISO works as expected.
How do I identify the cause of this behaviour, and, more importantly, make it go away?
[edit 1 below]
So I tried taking the selections from the manual install, filtered out anything that had disk|partm|grub and created a preseed partitioning file for the R320 from that; using that the machine never gets past partman: reading all physical volumes ...
[edit 2 below]
So .. not being able to find out why/how the provisioned system is never bootable with 16.04 I figured I'd try debian8 (Jessie) and ask them about preseed in IRC when it fails. Added Debian as an OS in foreman, associated it with the same preseed files as Ubuntu, changed the hosts OS from Ubuntu 16.06 to Debian, kicked off a build. Had to answer 2 installer questions (language, hdd for grub), and had a working server in about 30 minutes :/ Not what I was going for. Can anyone tell me how to find out why Ubuntu behaves so differently?
Cheers
How would I go about setting foreman up so it hands out a lease to any machine I connect to one of its interfaces using a private IP in the 192.168-range during provisioning, but to supply the client with it's actual IP (either public, or private in a 10.10 net) for operation via ansible?
I don't want to (it's also against our policy) to hand out DHCP on the production server network.
My foreman install is on ubuntu 14.04, running in a VM with two network interfaces; eth0 is static on the server network, eth1 is 192.168.0.1 in a 192.168.0.0/24 ... running in a VM.
We're a python shop, are using ansible, and are looking at integrating foreman with ansible to tackle bare-metal installs.
Is it possible at all to use foreman w/o puppet?
I found the documentation exhaustive but couldn't find a clear answer; the mention that "The Foreman installer uses Puppet to install Foreman." makes me think I will need to install puppet even though we don't intend to use it?
I'm playing with foreman (learning about both foreman and docker), and wanted to try the docker approach.
Working with an official container from docker hub
Can't find the docker file for it anywhere.
How do I find out which file gets executed when I fire up the container from within the container?
The reason that I'm asking is that I'd like the container to use both network interfaces of the host, one for the Web-UI, the other for dishing out DHCP and DNS on a separate private network.
If I execute the container with 'bash' tacked on I can pass my config options for dns and dhcp proxy network to it, but I lose all the other goodness, i.e. SSL certs, admin credentials, ...
I'm looking after a V40z (a loaner), running RHEL 4.8 (no, I can't update it, it's a loaner) that according to dmidecode has 8 DIMMs of 2048M each, which makes for 16GB RAM as far as I'm concerned.
The kernel, however, reports the total memory to be
cat /proc/meminfo
MemTotal: 14898844 kB
Now, on machines with integrated graphics chip-sets I've seen memory in the range of 128 to 512MB go "missing", but what would be an explanation for this huge discrepancy on a server? The way I see it it's missing 1.8G.