Cascading errors caused AWS to go down

http://status.aws.amazon.com/rss/ec2-us-east-1.rss

by Eliseann • 14 years ago

151 87 14 years ago

status.aws.amazon.com

eli• 14 years ago

In case your browser doesn't speak RSS:

Service is operating normally: Root cause for June 14 Service Event June 16, 2012 3:15 AM

We would like to share some detail about the Amazon Elastic Compute Cloud (EC2) service event last night when power was lost to some EC2 instances and Amazon Elastic Block Store (EBS) volumes in a single Availability Zone in the US East Region.

At approximately 8:44PM PDT, there was a cable fault in the high voltage Utility power distribution system. Two Utility substations that feed the impacted Availability Zone went offline, causing the entire Availability Zone to fail over to generator power. All EC2 instances and EBS volumes successfully transferred to back-up generator power. At 8:53PM PDT, one of the generators overheated and powered off because of a defective cooling fan. At this point, the EC2 instances and EBS volumes supported by this generator failed over to their secondary back-up power (which is provided by a completely separate power distribution circuit complete with additional generator capacity). Unfortunately, one of the breakers on this particular back-up power distribution circuit was incorrectly configured to open at too low a power threshold and opened when the load transferred to this circuit. After this circuit breaker opened at 8:57PM PDT, the affected instances and volumes were left without primary, back-up, or secondary back-up power. Those customers with affected instances or volumes that were running in multi-Availability Zone configurations avoided meaningful disruption to their applications; however, those affected who were only running in this Availability Zone, had to wait until the power was restored to be fully functional.

The generator fan was fixed and the generator was restarted at 10:19PM PDT. Once power was restored, affected instances and volumes began to recover, with the majority of instances recovering by 10:50PM PDT. For EBS volumes (including boot volumes) that had inflight writes at the time of the power loss, those volumes had the potential to be in an inconsistent state. Rather than return those volumes in a potentially inconsistent state, EBS brings them back online in an impaired state where all I/O on the volume is paused. Customers can then verify the volume is consistent and resume using it. By 1:05AM PDT, over 99% of affected volumes had been returned to customers with a state 'impaired' and paused I/O to the instance.

Separate from the impact to the instances and volumes, the EBS-related EC2 API calls were impaired from 8:57PM PDT until 10:40PM PDT. Specifically, during this time period, mutable EBS calls (e.g. create, delete) were failing. This also affected the ability for customers to launch new EBS-backed EC2 instances. The EC2 and EBS APIs are implemented on multi-Availability Zone replicated datastores. The EBS datastore is used to store metadata for resources such as volumes and snapshots. One of the primary EBS datastores lost power because of the event. The datastore that lost power did not fail cleanly, leaving the system unable to flip the datastore to its replicas in another Availability Zone. To protect against datastore corruption, the system automatically flipped to read-only mode until power was restored to the affected Availability Zone. Once power was restored, we were able to get back into a consistent state and returned the datastore to read-write mode, which enabled the mutable EBS calls to succeed. We will be implementing changes to our replication to ensure that our datastores are not able to get into the state that prevented rapid failover.

Utility power has since been restored and all instances and volumes are now running with full power redundancy. We have also completed an audit of all our back-up power distribution circuits. We found one additional breaker that needed corrective action. We've now validated that all breakers worldwide are properly configured, and are incorporating these configuration checks into our regular testing and audit processes.

We sincerely apologize for the inconvenience to those who were impacted by the event.

namidark• 14 years ago

Sounds like Amazon is doing something wrong, shouldn't it fail over to Battery then Generator?

WestCoastJustin3• 14 years ago

You're correct. It goes battery then generator. If they didn't use battery first, then when the power initial failed all systems would be off-line as the generator takes about 15-30 seconds to kick in.

rdoherty• 14 years ago

From what I can remember of a datacenter tour, most generators supply power to the batteries, which then supply power to servers. The batteries can only supply a few minutes (I think) of power, so the generators need to turn on immediately.

mrkurt2• 14 years ago

Battery backup isn't usually considered a "failover" step, just like your desktop battery backup doesn't actually do much other than stop charging when the power goes out.

Datacenters only really have battery backup to let the emergency generators come up.

ericabiz3• 14 years ago

The batteries at colocation facilities are only designed to hold power long enough to transfer to the generator. They're also a huge single point of failure. A better design is a flywheel that generates enough power. But datacenters are often hit with these generator failures (in my experience, once every year or so.)

Amazon had a correct setup--but not great testing.

By the way, these are great questions to ask of your datacenter provider: Are there two completely redundant power systems up to and including the PDUs and generators? How often are those tested? How do I set up my servers properly so that if one circuit/PDU/generator fails, I don't lose power?

There is a "right way" to do this--multiple power supplies in every server connected to 2 PDUs connected to 2 different generators--but it's expensive, and many/most low-end hosting providers won't set this up due to the cost.

(I ran a colocation/dedicated server company from 2001-2007.)

yaix• 14 years ago

"service event"

Added to my list of favourite euphemisms.

heretohelp• 14 years ago

1 in a (million/billion/trillion) I guess.

That'll make for a great horror story to tell though.

kbutler• 14 years ago

Isn't the moral of the story, "Check your backups"? There was a defective fan in one generator (sounds like it was findable via a test run?) and a misconfigured circuit breaker (sounds like it was findable by a test run).

Redundancy is only helpful if the redundant systems are actually functional.

olefoo• 14 years ago

Having been affected several times by colocation facilities bouncing the power during a test of the failover system, I can tell you that such tests are not without risk. Yes, you should test redundant systems, but how often, at what cost, and what risks are you willing to run while doing so.

It's a fact of life that when dealing with complex, tightly coupled systems with multiple interactions between subsystems that you will routinely see accidents caused by improbable combinations of failures.

gameshot911• 14 years ago

I agree, it sounds like this could have been discovered had the two backup power systems been properly tested.

Note how in the case of backup power, "properly tested" doesn't mean 'Does the generator turn on? Are we getting electricity from it? Ok, pass!'. It means running the backup generator in a way that is consistent with what you would expect in an actual power failure - i.e., for more than just a few minutes.

Same thing with storage backup. Checking your backups isn't just 'was a backup file/image created?', it means _actually trying to recover your systems from those backup files_.

patrickgzill• 14 years ago

The datacenter I use, has a policy to do full load runs for 30 minutes each month.

AWS would have found this, and been able to fix it in a timely fashion, if they did the same (the genset lasted for 10 minutes under load before failing).

jluxenberg• 14 years ago

"Those customers with affected instances or volumes that were running in multi-Availability Zone configurations avoided meaningful disruption to their applications"

"Meaningful disruption" is a bit of a weasel word; Amazon's own EBS API was down for almost two hours[1] despite being designed to use multiple AZs

[1] "the EBS-related EC2 API calls were impaired from 8:57PM PDT until 10:40PM PDT ... The EC2 and EBS APIs are implemented on multi-Availability Zone replicated datastores"

Guess the moral of the story is, if you require high availability then you must test your system in the face of an availability zone outage.

jrockway• 14 years ago

The RSS link was quite amusing. My Chrome instance downloaded the RSS file without displaying it. Then I clicked it to open, and it opened Firefox. Firefox showed its file download box, suggesting I open the RSS with Google Chrome.

Deadlock detected.

nowarninglabel2• 14 years ago

Does anyone know a solution for this? When I got upgraded to this version of Chrome (Version 20.0.1132.34 beta), it started downloading RSS feeds instead of displaying them. Much sadness has ensued :(

anaheim2• 14 years ago

AFAIK Chrome has never detected RSS correctly for me. I assumed this was special to Mozilla, with their "we should probably provide a basic version of everything - newsreader, FTP client, etc., even if it makes the browser a bit more bloated." Bit like emacs.

Chrome is a bit like vi, if you want more stuff, there's probably some sort of extension.

Sorry for the emacs/vi analogies, I'm not trying to flame :)

tedunangst• 14 years ago

On Linux, right? Firefox, or whatever gnome/dbus/opendesktop/gtk fuckery it uses has all sorts of strange notions about file types. When I download a tar.gz file, it saves a copy to /tmp, then launches a new instance of firefox with a file:// url, which opens a save file dialog.

dfc• 14 years ago

Respectfully, your setup is broken; not linux...

dkersten• 14 years ago

Firefox on Linux works fine for me for both RSS files and tar.gz files.

Osiris• 14 years ago

Opera detected the link as an RSS feed and asked me to subscribe to it with the built-in RSS reader.

I wonder if other browsers would benefit from being able to at least detect RSS feeds and display appropriate information to user.

hboon• 14 years ago

Safari does and pops up the default RSS client.

ars• 14 years ago

> I wonder if other browsers would benefit from being able to at least detect RSS feeds and display appropriate information to user.

Firefox does. It shows the feed right in the browser.

anaheim2• 14 years ago

TL; DR:

Shit happens. Don't use AWS as your only platform, you will get burned sometime. Guaranteed, you will also get burned if you try to host and run your own stuff. How competent you are determines which way you get burned less.

mhartl• 14 years ago

Or, if you use AWS as your only platform, accept that shit will happen from time to time. Unless your application is a matter of life and death, or unless billions of dollars are at stake, a little downtime now and then probably isn't that big a deal. (All my sites went down when Heroku did (including railstutorial.org, which pays my bills), but the losses are acceptable given the convenience of not having to run my own servers.)

rdl8• 14 years ago

I think it's reasonable to escalate criticism of Heroku for remaining in a single AZ. They have had plenty of time and resources to fix this, and haven't, despite being quite competent. I don't know if it is that they don't think it's necessary (due to the profile of their current customers) or what, but I wouldn't use Heroku for anything as long as they remain in a single AZ, and would be really reluctant to advise other people to do so. I obviously really like the Heroku team and product and would love to use them otherwise.

It wouldn't even need to be true seamless failover across AZs right away -- just offering a us-west and us-east Heroku would be enough for me, with shared nothing (maybe billing, or not even that), and then figure out redundancy yourself inside your app. Multiple regions is WAY better than multiple AZs within a region, too -- both for reliability and for locality.

Obviously a real seamless multi AZ/multi region solution would be much more technically impressive, useful to users, and Heroku-like, but they shouldn't let the perfect be the enemy of the good here.

spartango2• 14 years ago

While I'd agree with the general premise that diversification is a good thing in platform use if high-availability is a requirement, given that this outage was single-AZ, this particular outage should really highlight the point that your application should be multi-AZ scaled if it needs to be up.

scyscy• 14 years ago

Yes. Also, stop navel-gazing (usually that means stop reading Hacker News). Stop commenting on Hacker News as well. Funny thing about the Singularity/aliens/heaven--it'll come even if you don't spend a lot of time worrying about it.

mattdeboard• 14 years ago

What?

starship• 14 years ago

Actually, starting right now, AWS is probably your best bet.

Old story about Chuck Yeager from the 1950's: one time shortly after take-off, Yeager's aircraft suffered an engine failure, and he had to do an emergency semi-crash landing. When he realized that a mechanic had put the wrong type of fuel in the plane, he went looking for the guy. The mechanic profusely apologized, said he would resign and never work in aviation again. Yeager replied something along the lines of "Nonsense. In fact, I need someone to refuel my plane right now, and I want you to be the one to fuel it. That's because of all the guys here, I know you'll be the one guy who'll be sure to do it right."

Probably apocryphal, but the point has merit.

Monkeyget• 14 years ago

This is mentioned in 'How to win friends and influence people' where the anecdote is about Bob Hoover and Jet fuel put in a WW2 plane. It is used as an example that it is easy to criticize and complain but that it takes character to be understanding.

acdha2• 14 years ago

More accurately: “don't trust any single data center”. All of the people who complained were directly ignoring Amazon's own advice, not to mention decades of engineering experience.

Going multi-AZ, multi-region or multi-cloud will help, each step up that list being significantly more work for increasingly small returns.

jtchang• 14 years ago

Data Center Operator:

We've lost our main power. No problem though we have a backup generator so we are good!

... 5 minutes later ...

Uhh boss, our backup generator's fan crapped out. But no worries we have a secondary generator just for this kind of scenarion!

...10 minutes later and lights go out...

"Well damn...looks like we configured the breaker wrong. This is not a good day."

lutorm• 14 years ago

Things could be have been worse -- it could have been a nuclear power plant. oh wait...

forgotusername• 14 years ago

In my time at larger companies, DC power seems to be one of the weakest links in the reliability chain. Even planned maintenance often goes wrong ("well we started the generator test and the lights went out, that wasn't supposed to happen. Sorry your racks are dead").

Usually the root cause appears simple - a dead fan, breaker set to the wrong threshold, alarm that didn't trigger, incorrect component picked during design phase, or whatever else that gets the blame - things it would seem to a software guy that good processes could mitigate.

Can any electrical engineers elaborate on why power networks fail (in my experience at least) so frequently? I guess failure modes (e.g. lightning strike) are hard to test, but surely an industry this old has techniques. Is it perhaps a cost issue?

mrkurt2• 14 years ago

It's really incredibly complicated, and difficult to test fully. The bits of Amazon's DC that failed seem like stuff normal testing should catch, but the DC power failures I've dealt with in the past always had some really precise sequence of events that caused some strange failure no one expected.

As an example, Equinix in Chicago failed back in like 2005. Everything went really well, except there was some kind of crossover cable between generators that helped them balance load that failed because of a nick in its insulation. This caused some wonky failure cycle between generators that seemed straight out of Murphy's playbook.

They started doing IR scans of those cables regularly as part of their disaster prep. It's crazy how much power is moving around in these data centers, in a lot of way they're in thoroughly uncharted territory.

rdl8• 14 years ago

The even crazier thing is big industrial plants where they are using tens or hundreds of MW and have much lower margins than datacenter companies, so they run with dual grid (HV, sometimes like 132kV) feeds and no onsite redundancy. As in, when the grids flicker, they lose $20mm of in-progress work.

nl• 14 years ago

Rackspace has a pretty detailed report[1] of their 2009 outages, which is surprisingly similar to the Amazon problems.

[1] http://broadcast.rackspace.com/downloads/pdfs/DFWIncidentRep...

davps• 14 years ago

Power failures caused by lightning strike are relatively easy to test with platforms like RTDS [1] (I am not affiliated to RTDS).

I know that you can test in real time your electrical protection systems for almost all the possibilities you can imagine (thousands of them), for example: faults in your high voltage utility distribution system, breaker failures, coordination of the protection systems, lost of your back-up generator power. I don't know their systems or their philosophies, would be interesting for me to know why they don't parallelize groups of generators (at the backup system), so, when one generator fails, the power load are balanced to the others (and using well known schemes to avoid cascade failures).

[1] http://www.rtds.com/applications/applications.html

rdl8• 14 years ago

I assume you mean "Datacenter (conditioned) Power", not literally Direct Current power.

In my experience (in ~30 datacenters worldwide, and reading about more), the actual -48v Direct Current plant is usually ROCK SOLID, in comparison to the AC plant. It's almost always overprovisioned and underutilized, at least in older facilities, or those with telcos onsite (who, unlike crappy hosting customers, actually understand power).

My pro tip for datacenter reliability is to try to get as much of your core gear on the DC plant as possible -- core routers, and maybe even some of your infrastructure servers like RRs, monitoring, OOB management, etc. Ideally split stuff between DC and AC such that if either goes down, you're still sort of ok, or at least can recover quickly. DC and AC is even better than dual AC buses, since what starts out as dual AC can easily end up with a single point of failure later (like when they start running out of pdu space, power, or whatever), and dual AC is also more likely to have a closer upstream connection.

DC stuff is WAY simpler to make reliable and redundant, just uses larger amounts of copper and other materials.

Spooky23• 14 years ago

Not an EE, but I've observed a few things about electrical infrastructure:

- The work is usually done by outside contractors, working off of specifications that may or may not make sense.

- Some aspects of testing have the potential to be dangerous to the people doing them. (ie. if network switch fails in testings, no big deal. If some types of electrical switch break during testing, the tester is dead.) High voltage electricity is not a toy.

- IT and facilities staff usually don't talk much, and often don't understand each other when they do.

- There's no instrumentation. I get an alert when IT systems aren't configured right. Nothing from the other stuff.

- There is a wide variance in quality of electrical infrastructure that isn't obvious to someone who isn't skilled in that area. IT folks don't need to deal with computers built in 1970. Electricians deal with ancient stuff that may be completely borked all of the time.

damian2000• 14 years ago

I love this sentence: Those customers with affected instances or volumes that were running in multi-Availability Zone configurations avoided meaningful disruption to their applications; however, those affected who were only running in this Availability Zone, had to wait until the power was restored to be fully functional.

Translation: If you have a redundant (multiple-AZ) installation, then you were ok, if not then your server died.

drags• 14 years ago

Did anyone else run into issues with ELB during the outage? We're multi-AZ and could access unaffected instances directly without a problem, but the load balancer kept claiming they were unhealthy.

tzury• 14 years ago

Seems like deploying on two _physical_ regions (or more) is the best and only proven approach.

That could be within the global AWS, or even say, one cluster at AWS and the other at RackSpace/Linode, etc.

robryan• 14 years ago

Then you just need to worry about your application consistency with the replication lag. No silver bullet I guess.

mleonhard2• 14 years ago

I'm running https://www.rootredirect.com/ and http://www.restbackup.com/ in us-east-1, in multiple availability zones. Both sites remained up with no problems.

aparadja• 14 years ago

Does your rootredirect service actually attract paying customers? I'm genuinely interested to know.

joelcollinsdc• 14 years ago

Also, how do you do this without having to handle the customers DNS lookups as well?

mleonhard2• 14 years ago

The title is incorrect. It should say something more like "Cascading failures cause part of AWS to go down."

tysont• 14 years ago

On the plus side, the level of transparency that AWS displays and the detail that they provide seems above and beyond the call of duty. I find it refreshing and I hope that other companies follow suit so that customers can understand the details of operational issues, calibrate expectations appropriately, and make informed decisions.

rdl8• 14 years ago

They're less transparent and responsive than most datacenter or network providers -- it's just that most of those providers hide their outage information behind an NDA, so only customer contacts get it, vs. making it public.

smackfu• 14 years ago

Yeah, a good datacenter will have SLAs around the root-cause analysis document for any failures. Like a preliminary report within a day and a final report within 7 days.

acdha2• 14 years ago

I've also had a few cases where providers either outright lied or only gave details if you persisted in requesting them. Having to play that game gets old…

moe• 14 years ago

Can someone translate that to control rods and manifolds?

gosub• 14 years ago

Could it be possible to have power management the same way Erlang manages processes? Instead of 2 or 3 enormous backup power unit, hundreads of small ones to come in and out of use "fluently".

Cascading errors caused AWS to go down

Related Stories