The first thing you generally want to do when you have any new Storage system like SSD, Disk arrays or a Cluster Ceph, is benching. You will want to know how can read and write throughput. FIO is able to do that for you, here is an example: [global] ioengine=libaio invalidate=1 ramp_time=5 direct=1 size=5G runtime=300 time_based directory=/home [seq-read] rw=read bs=64K stonewall [rand-read] rw=randread bs=4K stonewall [seq-write] rw=write bs=64K stonewall [rand-write] rw=randwrite bs=4K stonewall You then will have a good output of everything you need to know.

Continue reading

I recently been faced on a classical problem on InnoDB which is the fragmentation, but on Galera. InnoDB engine doesn’t defragment on the fly and requires optimize maintenance sometimes to free disk space. But on Galera, which is a fault tolerance and high availability solution, it’s a problem having tables locked by an optimize procedure. Until Galera doesn’t support TokuDB and only fully support InnoDB, we had (with a colleague (Kevin aka Vinek)) to find a solution.

Continue reading

I recently heard of HSTS which is a way to force users to come back to your website in SSL if they’ve already be to HTTPS once. It is simple, just add this line: # HSTS (force users to come in SSL if they've already been once) add_header Strict-Transport-Security "max-age=31536000; includeSubdomains"; If you want to have an overview of a complete configuration with it, look at the my wiki.

Continue reading

My use case is specific but not isolated. When I’m at work, I’m connected to my VPN at home. I have a specific DNS at home for my domain in deimos.lan and this is very useful to avoid me to remind all the IP of the services I have. Sometimes, I want to connect to a home service from the VPN, but my bookmarked links have my home DNS which are unknown from the DNS at work.

Continue reading

And the thing has finally happened! In recent years,the idea of​​one day workingat Red Hatpleases me.As you may know, I’m working at eNovance and the news is: RedHat has acquires eNovance! So I belong now to Red Hat family :-). RedHat really wants to to have our OpenStack experience and I’m excited to work with other teams like Ceph (ex Inktank). I really hope it will be an excellent adventure and will enjoy it.

Continue reading

I really like this kind of news, new versions of docker and RedHat: Docker 1.0: this is the first stable version after more than one year of development. This version is production ready and I will start using it asap. RedHat 7: this is a good news as the version 6 started to be a little bit old now. What’s new? MariaDB replace MySQL \o/ Docker integrated systemd XFS by default Kernel 3.

Continue reading

The Percona replication manager (PRM) is a framework using the Linux HA resource agent Pacemaker that manages replication and provides automatic failover. This covers the installation of the framework on a set of servers. The PRM framework is made of 4 components: Corosync, Pacemaker, the mysql resource agent and MySQL itself. It’s easy to setup, better if you already know how to use Pacemaker and it works like a charm. In fact it setup a master and x slaves.

Continue reading

Author's picture

Pierre Mavro / Deimosfr


Qovery Co-Founder and CTO

Paris - France