“Glue records” are DNS entries in the top level domain name that allows for a domain to look at itself to resolve IP-addresses. In my case, using Lan2k.org as DNS, and wanting to host this myself, I needed to notify my registrar to add glue records into the “.org” -domain.
Run: host -t NS lan2k.org will say:
lan2k.org name server ns1.lan2k.org.
lan2k.org name server ns2.lan2k.org.
Do get this working, my registrar has added these entries as glue records into the .org-domain:
ns1.lan2k.org. 86400 IN AAAA 2a01:4f8:200:91f3::6
ns2.lan2k.org. 86400 IN AAAA 2a01:4f8:200:91f4::6
ns1.lan2k.org. 86400 IN A 18.104.22.168
ns2.lan2k.org. 86400 IN A 22.214.171.124
To read these records from .org, run this command to find which servers are hosting .org:
dig +short org. NS
Then ask one of those servers what records it has for lan2k.org:
dig +norec @b0.org.afilias-nst.org. lan2k.org. NS
This means I must have valid servers running on the IP’s listed above. If those server change IP, you need to notify the registrar so that they can update the glue.
Been forever since the last update. I’ve switched over from StartSSL to Let’s Encrypt for my certificate needs. It’s a project to get every webserver out there to use HTTPS. Check it out! I also want to share a quick reference on what ciphers to use for various servers. These can easily be found at cipherli.st.
The previous post mentioned that the machine would not start properly, and how to get a serial console going. The reason for the failure was that the NFS share did not mount properly. Not entirely sure why, but this solved it;
In /etc/fstab, I had:
[2a01:4f8:nnn:nnn::nnn]:/srv /srv nfs defaults,noatime 0 0
Changing to this seems to have solved it. I’m no expert in NFS, but it works for me:
[2a01:4f8:nnn:nnn::nnn]:/srv /srv nfs rw,noatime,hard,intr,rsize=8192,wsize=8192,timeo=14 0 0
Today I needed to see why a virtual machine under Xen was not starting up properly, and I knew it was possible to connect via a serial interface. Unfortunately this is not enabled by default (at least not in my installation). So here’s what to do.
ON GUEST MACHINE (my case is Ubuntu 12.04, using Startup system):
Add this in a new file /etc/init/hvc0.conf:
# Launch serial console for Xen
start on stopped rc RUNLEVEL= and (
container CONTAINER=lxc or
stop on runlevel [!2345]
exec /sbin/getty -L 9600 hvc0 xterm
# End of file
ON HOST MACHINE:
In the VM’s config (my case this is /etc/xen/some_guest.cfg), add a line like this: extra=”console=hvc0 xencons=tty”. That’s hvc-zero btw.
Now you have to reboot the machine, possibly shutdown and re-create the VM for the Xen config to kick in.
Now you can connect to the console by:
xm list (to find your VM’s ID)
xm console [ID] to connect. You hit Ctrl+] to escape from it, which on my Swedish Mac means pressing Ctrl+Å.
Looks like I’m running low on disk on my virtual machine (VM) running ownCloud. Luckily, there’s plenty more available on the host so today I’m expanding the logical volume (LV).
Step one – as always, BACKUP your data. I use rsnapshot together with ssh and mysqldump.
Step two – shutdown the guest.
Step three – On host run (as root): ssm resize /dev/vg0/cloudStorage -s+250G. Where vg0 is the name of your storage pool and cloudStorage is the name of your LV. SSM is a smart tool that takes care of partitioning, logical volumes, storage pools and file system.
Step four – Start your guest: virsh start ownCloud (or whatever the name of your VM).
Done. Happy clouding!