Web Hosting Talk : Web Hosting Main Forums : Dedicated Server : More than 5 hour downtime for Softlayer Cloud Computing

[gdtechind]

Today, The Cloudlayer Instance we are having with Softlayer has faced a major downtime. At the time of writing this post, the service is down (5.30 hour already).

The support staff hasn't provided much information about the reason for downtime.

They say, the power failed in server room #5 for Dallas. The power was restored soon, but i don't see the CloudLayer Instance coming up even after 5 hours.

Having account with Softlayer for more than 2.5 years, i have been very happy with their team and their network, but more than 5 hour downtime and still the support not offering any ETA, is making me think of shifting to some other provider.

I am not much aware of Cloud Computing Instances and how much time they would need to come back online. But, i don't think it should take so much of time.

I am still waiting for the service to be restored. I guess some other members on WHT may have also faced the trouble today.

Please feel free to add comments.

[UNIXy]

The power outage could have uncovered almost-dead hardware nodes, corrupted file stores, and erased non-persistently saved configuration settings. I find it interesting, but not surprising, how much overhead the "cloud" setup can bring about.

I'm sure SL is working around the clock on this. Hopefully all will be restored soon.

Regards

[JFSG]

Don't they have backup generators?

[web-1]

From this thread in the "Providers and Network Outages and Updates" section, the outage was only about 20 minutes. You didn't mention if you put in a trouble ticket or not. You should call them.

I've seen servers that power up but don't restart properly, and I've seen services like httpd not start after a power up.

http://www.webhostingtalk.com/showthread.php?t=894444

[gdtechind]

Originally Posted by web-1

From this thread in the "Providers and Network Outages and Updates" section, the outage was only about 20 minutes. You didn't mention if you put in a trouble ticket or not. You should call them.

I've seen servers that power up but don't restart properly, and I've seen services like httpd not start after a power up.

http://www.webhostingtalk.com/showthread.php?t=894444

I was probably the first person to spot the trouble as i was working on the server at that time. I created a ticket within 2-3 minutes of this trouble with their Cloud Computing Instance.

After that, i called them as well.

On the ticket there have been around 5 updates from LT Staff, but none of those provide any real information about what is happening and in how much time the things will be up to normal.

Simply, they are telling...

"We are waiting on our Information Systems team to bring the Cloud computing instances back online. Unfortunately we do not have a solid ETA on when they will be complete. We will update you when we have more information."

[webbysitedesign]

i also faced this outage

[gdtechind]

The instance is up after 10 hour of Downtime.
And, this happened first time in more than 2 years. And i can tell that Softlayer Support doesn't work good at all.

They are good if we need some hardware upgrade or we need to pay them additionally for any services.

Duing outage, the support is TERRIBLE.

[zubbab]

Maybe i have misunderstood something about the whole cloud hype but shouldn't it be possible for SoftLayer to just move your instance to other nodes that are up? I can't think that SoftLayer just have one server room where all their cloud nodes is..

[web-1]

Originally Posted by gdtechind

The instance is up after 10 hour of Downtime.
And, this happened first time in more than 2 years. And i can tell that Softlayer Support doesn't work good at all.

They are good if we need some hardware upgrade or we need to pay them additionally for any services.

Duing outage, the support is TERRIBLE.

2 years = 17520 hours

10 hours of that makes it close to 99.95% uptime.

Your math may vary depending on your mood

[gdtechind]

Originally Posted by web-1

2 years = 17520 hours

10 hours of that makes it close to 99.95% uptime.

Your math may vary depending on your mood

If it was 5 min downtime every month, i would find it acceptable.

If it is 10 hour in a day, I guess no one will be happy with it. And, i continued with them for this time, because they were really offering good services.

I am amazed, What SOFTLAYER MEANS BY CLOUD ???? If their team is seeing this thread, they should explain it. If cloud can have such downtime, i don't think they can ever be successful.

Are they testing the cloud services with their clients ?

[040Hosting]

As far as i know SL isn't even delivering cloud for over 2 years; more close to about 5 months (announced at 5-11-2009)

[rustelekom]

And also TS should be understand when he said - "10 hours of downtime unacceptable" he don't knew what happen. Yes,i can agree that when server died and need restart and i need wait upto 10 hours to get it back because tech sleep - it is unacceptable amount of downtime when in TOS provider promise me round of clock support.
But when cluster hard disk massive was died and amount of failed data big enough - tech need time to isolate issue, find solution and run restoration process, And it is only one scenario - solution may work or not and they need find another and so on.
It is hard to decide are they do the best or not without any real information but usually any host need save their client base and of course most of host value customer and especially within outage time. So, it is mostly trust question because not all host would like provide end-user with detailed information about issue.

[040Hosting]

Originally Posted by rustelekom

And also TS should be understand when he said - "10 hours of downtime unacceptable" he don't knew what happen. Yes,i can agree that when server died and need restart and i need wait upto 10 hours to get it back because tech sleep - it is unacceptable amount of downtime when in TOS provider promise me round of clock support.
But when cluster hard disk massive was died and amount of failed data big enough - tech need time to isolate issue, find solution and run restoration process, And it is only one scenario - solution may work or not and they need find another and so on.
It is hard to decide are they do the best or not without any real information but usually any host need save their client base and of course most of host value customer and especially within outage time. So, it is mostly trust question because not all host would like provide end-user with detailed information about issue.

That is a bit the problem with cloud computing; the larger your environment get; the larger the problems get when you are faced one.

While a cloud has many protections for problems; they will be rather serious when you really get one. Clouds are just as a VPS for me; something really fun to work with; but not yet for my production systems.

As you may have seen in other threads about this outage; most got back after about 20minutes, they are likely using dedicated servers.

[gdtechind]

I would say that they are not ready for Cloud Computing yet.

While i was having the downtime, i started digging the forums on Softlayer (some of them are restricted for hosting clients only, i guess.).

I found that one client on forum was telling that everyday there is around 5 - 10 minutes of accessibility problem with the Softlayer Cloud. The server doesn't go down, but isn't work fine during this period. After some time, things come back to normal.

I would say Softlayer is still at Amateur level for Cloud Computing solutions.

I prefer them as they can offer easy internal connectivity between my Dedicated Server and Cloud Computing Instance.

[Speckz]

Our CCI was down for almost 12 hours.