Monday, April 06

Geek

Daily News Stuff 5 April 2020

Monkey Business Edition

Tech News

  • How long does the battery on the new Ryzen-based Asus Zephyrus G14 really last?  (Tom's Hardware)

    For watching videos and browsing the web, 11 hours.

    For 3D rendering and gaming, about 2 hours and 20 minutes.

    That's not bad at all given its small size and high performance.


  • Mars needs COBOL programmers.  (Tom's Hardware)

    So do New Jersey and Connecticut.


  • Ha-ha, you fool! You fell victim to one of the classic blunders, the most famous of which is "Never get involved in a land war in Asia," but only slightly less well known is this: "Never roll your own crypto."  (Citizen Lab)

    Yes, you guessed it, it's our friends at Zoom again.


  • Don't bind your critical server applications to 0.0.0.0.  (CSO Online)

    Seriously, cut that shit out.

    Also, firewall everything. 

    Sometimes there's a mistake in the firewall rules, sometimes there's a mistake in the port binding, so always do both.

    Our current cloud server configuration has two firewalls for every server - the one provided by the platform and our own UFW / iptables configuration.  Which is sometimes too secure; we had problems with out new GitLab server because it was unable to make certain API calls to itself.


  • Stop reading this and update your Chrome.  (Forbes)

    Maybe I should have put that first?

    Nah.


  • I set up Gluster on my test cluster today.  It seemed the easiest option compared to Lustre and Ceph.

    It is indeed easy, and it works.  Performance is slightly odd, with one node running measurably faster than the other. 

    (It does recommend against running with just two nodes, because it can't form a quorum during a network partition event.  Not just in the docs, but when you run the setup command with the parameter replica 2.)

    I need to look at the SSL config settings to make it nice and secure, but a nice thing with Gluster is you don't need dedicated servers or block devices or even partitions: Just assign it a directory on an existing filesystem that exists on all the servers, and it will handle the rest.

    So you can easily use it on top of ZFS, for example, and take advantage of features like transparent file compression and deduplication.  And in theory snapshots as well, though I'm not sure how to restore a Gluster filesystem from a ZFS snapshot.  There's a page in the docs on running Gluster on ZFS that mentions restoring from snapshots, but only provides one sentence on how to do it, which doesn't really explain anything.

    Update: Aha.  It was under the "Features" submenu in the Administrator's Guide.

    Good: Gluster supports snapshots.
    Bad: It uses LVM.

    Hmm.


Video of the Day



This video is about as much fun as you can have and still accidentally prove Pythagoras's Theorem.


Disclaimer: You keep using that word.  I do not think it means what you think it means.

Posted by: Pixy Misa at 12:58 AM | No Comments | Add Comment | Trackbacks (Suck)
Post contains 489 words, total size 5 kb.




Apple pies are delicious. But never mind apple pies. What colour is a green orange?




50kb generated in CPU 0.013, elapsed 0.2015 seconds.
56 queries taking 0.1936 seconds, 338 records returned.
Powered by Minx 1.1.6c-pink.