justinsb

81 ms·

691.
by justinsb 15y ago
Thanks for the expert input. This is exactly why we don't have a SLA at FathomDB, because I don't think that they're worth the paper they're written on. It seems like you almost agree (?) Most cloud providers (including FathomDB) charge b
692.
by justinsb 15y ago
With physical hardware, the basic failure characteristics etc are well known. When the hardware is virtual or abstracted as it is on the cloud, you have no choice but to rely on the promises made by the provider. If you can't build a reli
693.
by justinsb 15y ago
Thanks for the good-luck wishes. Obviously it's a big undertaking, but we're seeing very promising results, so it's time to go all-in. Take that as you will :-) Incidentally, Oracle and IBM both have clustered database products, though th
694.
by justinsb 15y ago
I'm the author. I think you're conflating the SLA with the guarantees the cloud offers: one of the points I tried to make was that an SLA doesn't offer you any real guarantees; they only really benefit the service provider. If every SLA
695.
So long AWS, and thanks for all the fish: FathomDB (W08) focusing on its new DB(blog.fathomdb.com)
28 points by justinsb 15y ago|27 comments
696.
by justinsb 16y ago
FathomDB was also hit - 438012d3... was also my majorly disrupted AZ, 60dcfab3... was not significantly affected. Not sure about other us-east AZs. Thanks for pointing out that trick, by the way. Given AWS is exposing this anyway, they mi
697.
by justinsb 16y ago
If you're correct, that would be _very_ interesting. The AWS forums are full of statements to the contrary, I thought from AWS employees. The status updates never mention the particular zone (I presumed because the names don't match for d
698.
by justinsb 16y ago
It's a good analogy and I certainly accept your point. It could just be a marketing thing though: Suppose it's the same hard disk with a black sticker instead of a blue sticker. Drive with 1 yr warranty @ $100, 5 yr warranty @ $150, 20% a
699.
by justinsb 16y ago
I think it's great to check the SLA. However, there's enough wiggle room in the AWS SLA that I think this outage could continue for the rest of the month, and Amazon would still not owe a penny. I don't even know that the SLA covers this
700.
by justinsb 16y ago
It is possible. I think it's exceptionally unlikely. The one bit of information we have suggests that the root cause was a networking issue, which suggests SPOF.
701.
by justinsb 16y ago
Ah - sorry! I don't mean a legal contract, I mean more of a technical contract. e.g. "I won't pass a null pointer" style contract. In fact, the first bit you quoted provides an even stricter technical contract than the one on the main EC2
702.
by justinsb 16y ago
I have to disagree with you. The SLA is just a legal agreement that really serves to limit AWS's liability. Here's what the main EC2 page says: "Availability Zones are distinct locations that are engineered to be insulated from failures i
703.
by justinsb 16y ago
A quick tldr: Availability Zones within a Region are supposed to fail independently (until the entire Region fails catastrophically). Any sites that designed to that 'contract' were broken by this morning's incident, because multiple AZs f
704.
by justinsb 16y ago
And even if you had decided it was worth the cost, multiple AZs broke, which AWS promised wouldn't happen.
705.
by justinsb 16y ago
This is not the sites' fault. AWS promised that different Availability Zones would have no common single points of failure, but AWS broke that promise. I've written a blog post with more details: http://news.ycombinator.com/item?id=247189
706.
AWS is down, but here's why the sky is falling(justinsb.posterous.com)
347 points by justinsb 16y ago|80 comments
707.
by justinsb 16y ago
Don't let AWS hear that, or they'll charge us for their failures by rebranding it as a feature.
708.
by justinsb 16y ago
All well and good, but the elephant in the room is that multiple availability zones have failed at the same time. It looks like AWS have a single point of failure they weren't previously aware of.
709.
by justinsb 16y ago
AWS have now confirmed that this affects multiple availability zones. From the status page: "..impacting EBS volumes in multiple availability zones in the US-EAST-1 region"
710.
by justinsb 16y ago
AWS randomize the zones per account, so "your" -1b is not necessarily the same as "my" -1b. I'm only seeing problems in my -1c. Are you seeing 3 zones failing all under the same account? If multiple AZs are down, AWS are going to have som
711.
by justinsb 16y ago
Agree... I like what HelloFax is doing, and will definitely use it next time I have to deal with "one of those companies"
712.
by justinsb 16y ago
Because it's pretty easy to fake a traditional signature as well? Particularly if it only needs fax fidelity!
713.
by justinsb 16y ago
My understanding is that electronic signatures are actually legally binding, and an electronic signature just means anything that signifies intent to sign (e.g. no crypto required). So the patent office accepts your name in slashes, as I r
714.
by justinsb 16y ago
Without publishing the complete source for each test, these benchmarks are not very useful (the author does promise to publish sources for a new test in future). It appears that initially node.js was in 'development mode', so presumably th
715.
by justinsb 16y ago
Straight from the horse's mouth, here's OCZ's patch log for v129, which describes a fix for a software-induced data-loss issue: "Fixed rare corner case that could cause the drive to reset and clear user data" http://www.ocztechnology.com/f
716.
by justinsb 16y ago
Check out Facebook's FlashCache: they use SSDs as a cache on to traditional HDDs. You can have a read-cache to keep 'hot' data fast, and (if you want to) you can have a write cache that will buffer writes on SSD before flushing to HDD. I d
717.
by justinsb 16y ago
I think there's a bigger problem here - this isn't just normal bathub-curve component failure. What seems to be happening is that the firmware panics due to an assertion being hit, and what could be a hiccup that requires a reboot becomes
718.
by justinsb 16y ago
Sure, just sharing my personal experience. The pathological failure behaviour due to a software bug is not something I'd previously considered when choosing a drive. I hope you've got good backups.
719.
by justinsb 16y ago
Avoid OCZ. I had a Vertex 2 SSD fail after about two weeks. What's really a big gotcha is that it fails 100% immediately; there are no warning signs as there often are with a hard disk. I looked into it: I believe that there's actually a
720.
by justinsb 16y ago
Jeff Barr's post: http://aws.typepad.com/aws/2011/02/cloudformation-create-you...

More ›