docker-registry behind a few proxies

Running a private docker-registry behind a few proxies took me while to configure, because I had several things that I couldn’t move. In particularly, it is an nginx in front of everything, and the docker-registry that I wanted as a “real” service, because I am still learning the docker ways, and I don’t want it as a container, yet.

I installed the docker-registry in a KVM VM, on a CentOS 7 – a standard business requirement one might say.

That part is easy: fetch the virt7-testing repo, as described here, and move on: Continue reading

Confluence halts, JIRA works normally after replication set-up?

After setting-up replication, if JIRA starts just fine, but Confluence freezes with something similar to:

You cannot access Confluence at present. Look at the table below to identify the reasons.
Type Description Exception Level Time
cluster Non Clustered Confluence: Database is being updated by another Confluence instance. Please see http://confluence.atlassian.com/x/mwiyCg for more details.

Your server id is: SOME-CODE-GOES-HERE

fatal 2015-04-27 00:25:15

This page will automatically update every 60 seconds.

add

binlog_format=row

to your server.cnf, as described here.

Confluence and JIRA memberOf problem against OpenLDAP

I’ve just spent hours trying to figure out filters for either users or groups that would allow Confluence and JIRA to authenticate against OpenLDAP only those users that had “member” attribute for respective groups, all in vain. Both Confluence and JIRA simply ignore group membership unless “memberOf” attribute is used during the search. But simply turning the point around works – do not try to force Confluence or JIRA to use “member” attribute found in groups, but simply add “memberOf” attributes to each user you’d like in respective groups.

openSUSE 13.2 x86_64 froze in yast

An openSUSE 13.2 workstation I use for some time now simply froze on clicking “Yast >> System >> Bootloader >> Kernel_parameters (tab)”.

Screenshot from 2015-04-19 23:32:32

The only thing I can find on that topic is this post, descriptive, but not particularly useful in tracking the problem. As said there, a reboot fixed the things up.

Have you tried to turn it off, and on again?” is still a preferred method in IT?

mariadb 10 multi-master multi-database replication

Unlike mysqld 5.X, mariadb 10.X offers multi-master multi-database replication. That means it can use a single mariadb-server 10.X instance as a replication slave for all the other mysql-servers we need. Unlike supporting this combination, the very installation and set-up is rather straight-forward: Continue reading

Docker behind firewalld

Running Docker containers behind a firewalld can be a routing nightmare. I had to use CentOS 7 docker images on a customised CentOS 7 host, and the situation turned into an incompatibility fest pretty soon after I figured out the followng:

  • CentOS host came with no firewall, and systemctl listed dbus-org.fedoraproject.FirewallD1.service,
  • Dockerised CentOS containers have no systemd,
  • Docker’s internal routing isn’t exactly the shiniest piece of documentation on Docker,
  • IPTables-services and firewalld shouldn’t work simultaneously, and usage of IPTables-services is strongly discouraged on new hats, in favour of new the interface – firewalld,
  • Docker’s daemon uses own interface to write to Netfilter, that can be clearly visible by an “iptables -L” inspection,
  • Docker (apparently) creates random RFC1918 addresses for new containers,
  • Docker assigns two IPs for each container regardless of the third IP you might call for on the command line during “docker run…”.

Fun, eh?

After a trillion of attempts, here is the most sane and simple solution I have come by for now: Continue reading

Re-install pkg for the win

This morning another pkg update && pkg upgrade failed on my FreeBSD 11 CURRENT amd64. It was some 80-ish packages, and it wouldn’t go past py33-atspi-2.12.0_1. It took a make deinstall && make reinstall of the /usr/ports/ports-mngmt/pkg to complete the upgrade.

I couldn’t figure out in details what caused this awkward situation.

Fortunately, with the exception of the necessity to keep enchant locked if I want gedit to have language dictionaries available in the spell-checker plugin, this is the first clumsiness of the pkgng tool I’ve run into in several months of usage, even under the CURRENT branch.

There, just a brief note! 🙂

The conjugate function for clang 3.4.1 on FreeBSD fails?

Dealing with the conjugate of complex numbers brought me some linker failures for clang 3.4.1 on FreeBSD 11 CURRENT amd64:

vanja@current:/tmp % cat konjugat.c
#include <complex.h>
#include <stdio.h>

int main(void)
{
double complex kpx = 1.0 + 3.0*I;
double complex konjugat = conj(kpx);
printf(“Kompleksni broj je %.2f%+.2fi\n”, creal(kpx), cimag(kpx));
printf(“Konjugat od kpx je %.2f%+.2fi\n”, creal(konjugat) ,cimag(konjugat));
return 0;
}

Continue reading

enchant stole the languages in gedit-plugins

A note on strange gedit behaviour: after the last update of ports  – yes, there is always something wrong with ports after an update, no matter what it looks like – my gedit wasn’t able to do spell checking, though I was sure I had hunspell and aspell for at least English and German installed system-wide.

Gedit simply displayed no languages in the appropriate menu.

It would re-compile and re-install even from ports, just fine, but no re-compilation of plugins, python bindings or dictionaries helped until I re-installed enchant from ports! There, it might help someone figure out where have all the languages gone:

pkg info gedit\* \*enchant \*aspell
gedit-3.14.2
gedit-plugins-3.14.1
enchant-1.6.0_4
py27-enchant-1.6.5_6
aspell-0.60.6.1_5
de-aspell-20030222.1_1
en-aspell-7.1.0_1
it-aspell-2.2.20050523.0_1,2