rbranson

24 ms·

241.
by rbranson 13y ago
+1 I don't think he's upset about LWW being characterized as a weak safety constraint, but that the perception that what's provided by Cassandra is equivalent to per-key LWW. While it doesn't serve to completely eliminat
242.
by rbranson 13y ago
Agreed in general. I think the statement that "you will lose data" is a bit simplistic. Given enough chances, all systems will lose data. One could make a pretty strong argument that the default LWW approach used by Riak and Vol
243.
by rbranson 13y ago
Reads in Cassandra will never trigger a compaction. Compaction is only be triggered by writes. If the compaction queue gets behind due to overloaded CPUs / disks, reads will begin to slow down.
244.
by rbranson 13y ago
This is highly unlikely. I'm on a team that administers an EC2 infrastructure with tens of terabytes of RAM, and I've never seen anything that resembles memory errors. Considering the findings of http://www.cs.toronto.e
245.
by rbranson 13y ago
We are still a very serious user of Redis, and I agree with this. Redis works well as a persisted networked heap, which very different from a database. IMHO Redis would have to trade off a lot of it's value to become a "real datab
246.
by rbranson 13y ago
I assure you it did not.
247.
by rbranson 13y ago
Just to be clear -- this doesn't mean we have eliminated Redis. It's still a very useful tool.
248.
by rbranson 13y ago
This is really a classic criticism of capitalism clothed in a Dropbox t-shirt. http://en.wikipedia.org/wiki/Criticism_of_capitalism#Ineffic...
249.
by rbranson 13y ago
First, let me concede that Cassandra has had a storied history of terrible read performance. However, if the last time anyone looked at Cassandra for read performance was 0.8 or used size-tiered compaction, I'd encourage them to take anothe
250.
by rbranson 14y ago
Can you cite anything for this? Wear leveling is not the same as log-structured storage.
251.
by rbranson 14y ago
Can you provide an example of where Riak was "nearly impossible to manage operationally without dedicated staff at scale and the tools to introspect and analyze and deal with failures aren't robust enough yet"?
252.
by rbranson 14y ago
Did you even read what aphyr wrote before posting this? It detailed that unlike HTTP, which in general supplies one request per message (GET, POST, etc), a Riemann message contains potentially hundreds of different requests that must be pro
253.
by rbranson 14y ago
This can be considered independent of the fact that the income increases due to the gains in productivity (inflation-adjusted GDP/capita) largely accrued to the top earners, and not across the entire income spectrum.
254.
by rbranson 14y ago
The statistic is just to show the growth in our data volumes since last year. As with most social sites (and the web in general), the number of writes is a very tiny slice of the I/O pie.
255.
by rbranson 14y ago
The short answer is we don't do it very often, and it's often complex (data sizes, sharding, and availability concerns), so we do it on an adhoc basis and in a semi-automated fashion.
256.
by rbranson 14y ago
This isn't clear from the blog post, but we only use autocommit for read-only queries. Using autocommit is only really a major boost if you have a very high skew of reads to writes like we do (50:1).
257.
by rbranson 14y ago
While this is certainly useful, it's pretty basic. Something that I find is a much closer fit to UNIX pipes is iterators. UNIX pipes work similarly in that all of the commands in the pipeline are executed in "parallel" and the OS passes dat
258.
by rbranson 14y ago
Note that the State of California's GDP is more than 6x that of Greece, while only holding roughly half the absolute amount of debt Greece does.
259.
by rbranson 14y ago
By that definition data containers (dictionaries, lists, etc) built into popular languages or standard libraries are databases.
260.
by rbranson 14y ago
memcached is a database now? IMHO Redis barely registers as a database; it's more of a shared heap that supports snapshotting. At Instagram we use PostgreSQL, Redis, and Cassandra.
261.
by rbranson 14y ago
Even if you ignore all the externalities, hydro-fracking is still much more expensive than traditional shallow well land drilling operations, and it requires higher oil prices to sustain them. The 8 massive oil fields in Saudi Arabia (as we
262.
by rbranson 14y ago
I want to preface my comment: this is impressive work, congratulations on shipping, and this is what MongoDB should have been from the start. In reality, most transactional database deployments are heavily skewed towards read workload, so r
263.
by rbranson 14y ago
"Who can see my web profile? If your photos are set to public, anyone will be able to see your profile by visiting instagram.com/[your username] on the web. You do not have to be an Instagram user to view a public user’s profile on the web.
264.
by rbranson 14y ago
Wrong. The designer of web profiles (Maykel Loomans) joined Instagram before the acquisition and this has been in the works for many months.
265.
by rbranson 14y ago
Imagine the horror when you find out all the worldly financial assets you've accrued over your entire life are just a few database entries dispersed over a few databases on some spinning disk in some rack in some datacenter somewhere?
266.
by rbranson 14y ago
AWS actually operates 5 independent data centers (called availability zones) within the same complex that make up the us-east-1 region. This makes them significantly more likely to experience an outage in 1 availability zone, but also makes
267.
by rbranson 14y ago
Actually this is much closer to theft, because it transfers value without consent. Copyright infringement does not transfer value.
268.
by rbranson 14y ago
I'd love to have actual distributed transactions that could scale indefinitely and not create availability issues. We actually get a steady stream of user complaints about inconsistencies between counter caches and what appears in results.
269.
by rbranson 14y ago
Given the context of the OP, a brand new project or startup, my advice still stands true.
270.
by rbranson 14y ago
This is why the "SQL Cookbook" is good, it has dozens and dozens of examples with permutations in Oracle, MS SQL, MySQL, and PostgreSQL.

More ›