geertj

15 ms·

181.
by geertj 4y ago
Any such system would probably use DC, which would be much more compact than what you describe.
182.
by geertj 4y ago
I wish the Oxide web site had a little more detail on the software running on this. Does This use a custom hypervisor? If so what is it based on, and what are its capabilities?
183.
by geertj 4y ago
The sibling comment is correct. The EFS mount helper starts up and manages an stunnel process. We have not seen a significant impact on latency from the stunnel process.
184.
by geertj 4y ago
> Under NFSv2 and NFSv3, the numeric user and group id is used to determine permission, and these must be aligned between the client and server. Technically the server doesn't need to have a UID/GID database that's aligned
185.
by geertj 4y ago
We currently don't support Windows as a client. The main reason is that the builtin Windows NFS client is v3, while EFS is v4+. Some people have reported success mounting EFS on Windows using a third party NFS client, but I cannot comm
186.
by geertj 4y ago
> It looked like work on this had stopped. Is there still hope that it might become a published RFC? I don't know, I hope it will. Not to go on too much of a tangent, and at the risk of sounding like my employer's fanboy, but
187.
by geertj 4y ago
> Will EFS be updated to use the NFS-TLS RFC once it settles down some? I can't commit on a public forum for obvious reasons but we'll definitely take a serious look at this, especially when the Linux client starts supporting t
188.
by geertj 4y ago
As others have said I think the title was more clickbait-y. Olaf was the person who wrote the in-kernel NFS server so I think he at least appreciated NFS somewhat. The paper makes a few reasonable criticisms many of which are addressed now.
189.
by geertj 4y ago
> I'm not sure how common this extension is, at least the Linux server and client support it out of the box. Isilon also supports it, but it must be explicitly enabled. It is very common. I’m not aware of a v4 server that does not s
190.
by geertj 4y ago
(PM-T on the EFS team) S3 and EFS actually are quite different. Files on EFS are update-able, rename-able and link-able (I.e what’s expected from a file system), while S3 objects are immutable once they are created. This comes from the unde
191.
by geertj 4y ago
> They forced all car manufacturers to transition to EVs. That happened after Tesla had proven the model though, primarily that you can build an EV that has sufficient range and that people find desirable.
192.
by geertj 4y ago
Thanks for digging up the original quote and clarifying what I was replying to. Agreed, income per hour indeed doesn’t depend on competition until the next adjustment.
193.
by geertj 4y ago
> The only reason competition matters at all is because of the difficulty adjustment. Which seems relevant here (I think you’re both right?) because with more competition the next adjustment will be very quick, and will adjust your expec
194.
by geertj 4y ago
My style as well. I would add one thing which I call “critical dimension analysis.” Try to identify the main customer driver for the domain you’re in (could be cost/performance/something entirely different), and invest at little m
195.
by geertj 4y ago
> The impact Tesla has on climate change is so small as to be complete negligible. How are you defining impact? I think it's reasonable to conclude that Tesla caused the global move away from IC cars to EVs. According to the EPA, 27
196.
by geertj 4y ago
> Producers seek the profit maximizing price, not the highest price (those aren’t the same). Is this true when you only have fixed amount of the commodity? Wouldn’t players try to optimize the total NPV of all resources sold into the fut
197.
by geertj 4y ago
I came here to write the same thing. Things I had hoped this would go into is how goroutines grow their stacks, and how they are preempted.
198.
by geertj 4y ago
Not sure if it’s concerning but it caught my eye as well.
199.
by geertj 5y ago
> There are 2 backlogs, one run by product and one run by dev/ops/arch/sec. Over the course of a PI (about 1 quarter), capacity planning usually allocated 70/30 between the two backlogs. Does "run by" mean t
200.
by geertj 5y ago
Hi! Unless you need to drive that 100 MB/s continuously, you can use the bursting throughput mode, which means you don't have to separately provision and pay for throughput. If you do need to drive 100 MB/s continuously, then
201.
by geertj 5y ago
> Is there an updated version of latency.txt? Would be interesting to compare against. Our updated docs [1] contains a latency table. The official answer is that previously our read latencies were low single-digit millisecond and now th
202.
by geertj 5y ago
> Also, does today's update apply to GovCloud regions as well or just Commercial for now? It also applies to GovCloud.
203.
by geertj 5y ago
> Could you use EFS to host a horizontally scalable relational database by using finely sharded SQLite dbs? Like if you had 1 db per user, for example. This is actually a more common use case than I had imagined just one year ago. It wo
204.
by geertj 5y ago
There's different schools of thought around this and I certainly understand your perspective. At AWS, carrying a pager at limited times (in our team, 2-3 weeks per quarter as mentioned in the link) is considered an important part of ou
205.
by geertj 5y ago
NFS workloads are typically metadata heavy and highly correlated in time, so you can achieve very high hit rates. I can't share any specific numbers unfortunately.
206.
by geertj 5y ago
For ML training, where you repeatedly open the same files over and over again, the NFS 'close to open' coherency protocol (where every open() and close() is a round trip) is not a good fit. Once we support NFS delegations, this wi
207.
by geertj 5y ago
(PMT on the EFS team). Yes, the wordings are carefully formulated as they have to be signed off by the AWS legal team for obvious reasons. With that said, this update was driven by profiling real applications and addressing the most common
208.
by geertj 5y ago
> Am I on the right track? :) The caches are local to each AZ so you get the low latency in each AZ, the other details are different. Unfortunately I can't share additional details at this moment, but we are looking to do a technica
209.
by geertj 5y ago
Hi, thanks for the link! I just tried this on EFS but it doesn't give the expected answer. The program uses NFS NULL calls. I'm guessing we haven't optimized those yet given that they are typically only used at mount time, bu
210.
by geertj 5y ago
I'm the PMT for this project in the EFS team. The "flip the switch" part was indeed one of the harder parts to get right. Happy to share some limited details. The performance improvement builds on a distributed consistent cac

More ›