paulfurtado

8 ms·

31.
by paulfurtado 5y ago
Slightly pedantic: ec2 doesn't actually support nested virtualization on any instance type I know of, but does have baremetal instance types that support virtualization. The reason I mention this is because, sadly, baremetal instance t
32.
by paulfurtado 5y ago
The fuses are typically 10A and there are products that go right up to that. You can probably safely use 8A in most cars. So about 100 watts.
33.
by paulfurtado 5y ago
Is this really true? As far as I know you can be perfectly compliant with EU law by running in AWS's EU regions.
34.
by paulfurtado 5y ago
It's not usable for cross-domain AJAX requests, is implicit, and it allows for drive by request attacks that must be mitigated with csrf token headers. Using token auth headers can avoid the use of csrf tokens and make your backend onl
35.
by paulfurtado 5y ago
Unix sockets only require as much permission as creating a file in a directory. If a program has write access to a directory, it can create a unix socket. File permission on the socket then dictate which users may connect to it.
36.
by paulfurtado 5y ago
That still significantly raises the barrier to entry until someone writes a generic CSI implementation that translates to the flexvolume API. And then it's still yet another daemon running on every kubernetes node 24/7. Regardless
37.
by paulfurtado 5y ago
Sad to see flexVolume deprecated. CSI is a more powerful interface and I knew this was coming, but the simplicity of the flexVolume exec model for node-local volumes was an amazing feature. With very little effort, you could use them to cre
38.
by paulfurtado 5y ago
On linux (and I think even osx), even when stdin, stdout, and stderr are redirected, you can always get at the current session's tty via "/dev/tty" unless the executing program spawns a new session which would be ve
39.
by paulfurtado 5y ago
At this point you can consider all non-nitro instance types to be legacy: it's been years now since all new instance types have been nitro and pretty much every instance family has a nitro replacement that is both faster and cheaper th
40.
by paulfurtado 5y ago
Windows may actually let you override the EDID info, it's been a while since I used windows but I think it lets you do it in device manager... from the era when displays came with drivers on a cdrom. There are also EDID cloner/emu
41.
by paulfurtado 5y ago
Problem with Gira is I've never actually figured out how to purchase it as a consumer. The products look amazing, but it seems impossible to buy in the US.
42.
by paulfurtado 5y ago
FWIW, RPM based distros can use delta RPMs
43.
by paulfurtado 5y ago
This is a controller we wrote in order to make local storage actually usable. Internally we have a "kube cluster operator" that handles the cluster-level ops like this, gracefully draining nodes, etc.
44.
by paulfurtado 5y ago
Cool thing about finalizers is that there's no straightforward way to delete them from the CLI, you need to explicitly patch the object to delete the finalizer so that goes a long way towards preventing instant mass deletion, they real
45.
by paulfurtado 5y ago
No database should lack an unsynchronized replica in general. A kubernetes admin could accidentally drain the host, but we run in AWS and AWS can kill nodes a lot faster than we can haha. We have enough servers in AWS that at least 5 fail p
46.
by paulfurtado 5y ago
We do actually automate this away by deleting PVCs when they are tied to PVs on nodes with our drain taint on them and their pod is in the pending state. We also delete pods owned my StatefulSets that have missing PVCs. These two control lo
47.
by paulfurtado 5y ago
Say you have 100 database clusters and you want to schedule them on dedicated servers. A kubernetes node can advertise those local block devices as persistent volumes. Your StatefulSet can request that class of local storage and the kuberne
48.
by paulfurtado 5y ago
Vitess can be thought of as a very smart MySQL proxy with additional MySQL management features built in. Centralized vtgates proxy to vttablet services which run on the same host as MySQL, the tablet then queries the local MySQL server. On
49.
by paulfurtado 5y ago
You can be fairly good about it, but there's always state with the potential to be left behind or become inconsistent: transitive dependencies of packages sticking around, packages with pre/post scripts, services writing out data
50.
by paulfurtado 5y ago
Transmission also does have transmission-cli
51.
by paulfurtado 5y ago
Kubernetes implements load balancing via iptables probability rules. It creates "magic" static IPs on all kube nodes for ClusterIP services and makes iptables randomly select backend pods, making for a highly efficient zero-usersp
52.
by paulfurtado 5y ago
cillium reimplements kube-proxy's functionality. There may still be a couple of features missing, but it can definitely take over service routing more efficiently. I haven't checked in a bit but I'm pretty sure it covers Node
53.
by paulfurtado 5y ago
As someone who hates making frequent grocery trips and drinks milk too intermittently to ever finish a normal bottle before it goes bad, I highly recommend buying UHT (Ultra High Temperature) pasteurized milk which does actually often have
54.
by paulfurtado 5y ago
If running as a privileged user, you can run the process in just a PID namespace with no other container features enabled using the unshare command. Systemd can also handle it by killing the processes by cgroup i believe, and for non servic
55.
by paulfurtado 5y ago
It's been a while, but gnome-terminal on earlier gnome 3 releases used to segfault very frequently for me. It's the thing that finally made me seek out other terminal emulators
56.
by paulfurtado 5y ago
Doing it with eBPF gives you a hook for each retransmit, this makes it possible to know the exact connection, process, and network interface that hit the retransmit and allows you measure things like "100s of hosts are hitting 1000s of
57.
by paulfurtado 5y ago
That said, the ABS tree and the AUR in arch, combined with the general avoidance of patches, do make arch one of the easiest places to play with bleeding edge software. With the toolchain in place, so many packages are just a one line versi
58.
by paulfurtado 5y ago
Package versions don't remain in the repo for long so if you don't update the package index you'll find that many packages are not installable. There isn't really a snapshot of compatible versions at a given point in tim
59.
by paulfurtado 5y ago
> More often, sure. But for software delivery from the production side, the common viewpoint seems to be that deploying more often leads to less pain in total, because you're making smaller increments so the individual failures are
60.
by paulfurtado 5y ago
Yeah, definitely usually fine to avoid kernel upgrades, so that is a saving grace. The issue is just that pacman -Syu is going to be full of surprises

More ›