dikei

16 ms·

121.
by dikei 3y ago
"version + hash" is ugly though. I trust the publisher of my base image to keep compatibility even if they update their image and trust my test suites to detect any issues, so I just use version without the hash nowadays.
122.
by dikei 3y ago
AFAIK, C-A-D doesn't work over ssh.
123.
by dikei 3y ago
So Confluent pulled a "ship of Theseus" on Confluence Cloud, it still looks like Kafka, but everything underneath has been replaced by a proprietary system.
124.
by dikei 3y ago
I've encountered issue when NFS mount can block regular system reboot when the network connection is broken.
125.
by dikei 3y ago
One the most important thing about SQL is learn to read the output of EXPLAIN queries: it's the key to understand why your queries are not running as fast as you'd like.
126.
by dikei 3y ago
There can be consistency issue if you run multi-cluster BigTable, because multi-cluster replication is asynchronous.
127.
by dikei 3y ago
Unless you're doing multi-cluster replication, there is no log shipping in BigTable: the data replication within a cluster is taken care of by the underlying filesystems. Single-cluster BigTable is strongly consistent.
128.
by dikei 3y ago
Lima has experimental supports for Apple's native virtualization framework, but the default is still QEMU.
129.
by dikei 3y ago
Fun stuff: "What time is it ?" is a question that a vector clock can't answer, since it does not provide total ordering.
130.
by dikei 3y ago
Ha, it's the reverse for me. I used to use Debian base image, but now switch to Ubuntu LTS base images almost exclusively due to them offering newer versions of most core libraries, and longer support duration.
131.
by dikei 3y ago
If you mock out one end of the test, can it really be considered end-to-end anymore? Isn't it just UI testing now ?
132.
by dikei 3y ago
Exactly, it's similar to Linux live kernel patching. It's certainly desirable, but with the support of modern orchestration platform, most of the time it's simpler to migrate services off the nodes, update, reboot then redepl
133.
by dikei 3y ago
I mean he's gotta be an excellent mental gynmast to speak about real, profitable, value-adding businesses and crypto in the same post.
134.
by dikei 3y ago
Yup, and Apple SoC already have the DRAM packaged together with the CPU, only the SSD is left out now.
135.
by dikei 3y ago
> One of the selling points, which is now understood to be garbage, is that you can use different databases Yup, and even if the ORM supports multiple databases, once you have committed to one database, it's almost impossible to swi
136.
by dikei 3y ago
> you definitely cannot rely on clocks to determine (reliable) ordering across machines You can if you consider "unknown/possibly concurrent" also a valid outcome: for any 2 events A and B, True Time can definitely answers
137.
by dikei 3y ago
This looks pretty good, similar to Designing Data-intensive applications, but with more focus on Distributed Systems.
138.
by dikei 3y ago
You can build system that do not require physical clock synchronization, but using physical clock often lead to simpler code and major performance advantage. That's why Google built True Time, which provides physical time guarantee of
139.
by dikei 3y ago
> There's multiple overlapping standards, each with a dozen different variants. The Keycloak OIDC config refers to OAuth2 over a dozen times. To put it simply, OIDC = OAuth2 + id_token (a standard way for the provider to let your cl
140.
by dikei 3y ago
> I've never used it, but they can ban your app token, right? So I assume you need to make sure you're keeping your own up-to-date copy of all claims you have access to. That's why you should always provide multiple method
141.
by dikei 3y ago
The other stuff I do is to write YAML, HCL, and other type of configuration files.
142.
by dikei 3y ago
True, I was only talking about Kernel-Userland. It's also why closed-source software often include every dependencies in their installation package. And even that is sometimes not enough, like when the host's glibc is too old.
143.
by dikei 3y ago
> In practice, they compile only with specific versions of GCC, CMake and libc That's pretty easy nowadays thanks to container. You don't have to muck around with environment variables to get the build system to detect your cus
144.
by dikei 3y ago
Linux Kernel userland ABI/API is very stable, and backward compatible: you can take a binary compile for Linux 1.0 and run it today. The same applies for CPU instruction, backward compatibility means stuff targeting i686 instruction se
145.
by dikei 3y ago
> Relying on anyone else to provide a binary categorically disqualifies someone as a systems administrator. Come on, that's too harsh. We're not all Gentoo users, most sysadmins use packages provided by the distros. There'
146.
by dikei 3y ago
I use both, though I still prefer JetBrains IDE for coding.
147.
by dikei 3y ago
Fleet has been a multi-year effort, gotta applaud JetBrains for willingly take the bet and continuously investing in developing its products. It's way easier to rest on their laurels and only provide maintenance updates to IntelliJ.
148.
by dikei 3y ago
Nitpick: it's CO2 scrubbers, not O2. A scrubber primary mission is to absorb CO2, not generate O2.
149.
by dikei 3y ago
Agree, PostgreSQL enum is really just for mapping with an enum in application code. And just like you should not remove an enum or change their ordinal in application code, don't do that to a PostgreSQL enum
150.
by dikei 3y ago
.NET is a managed-language with a VM. In such language, a memory error in managed-code will often trigger a jump back to the VM, where they can attempt to recover from there. For native code, there's no such safety net. Likewise, even

More ›