CloudatCost Issues
-
@scottalanmiller said:
One of the two lab servers that had to be reimaged worked. The other looked like it worked, I could see it via the console, but the IP address didn't work and the password didn't work. So reimagining AGAIN. Two days now. I see where Aaron was having problems. With the imaging taking a full day and then failing more often than not, this really does get painful.
I better not have to reimage mine. That'd be a pain in the ass.
-
Just received this via email:
Cloud At Cost System Message:
03/19/2015 10:30 AM
CloudAtCost will be performing storage maintenance starting tonight. This maintenance will be for the purpose of adding additional storage at our Kitchener datacenter to accommodate our overwhelming growth. This maintenance will resolve disk io performance, occasional power on issues, server hangs and other performance issues.We will begin tonight at 8PM EST and will continue through the weekend.
NOTE: No downtime will be expected.
If you have any questions please let us know.
CloudAtCost Team.
-
@Danp said:
Just received this via email:
Cloud At Cost System Message:
03/19/2015 10:30 AM
CloudAtCost will be performing storage maintenance starting tonight. This maintenance will be for the purpose of adding additional storage at our Kitchener datacenter to accommodate our overwhelming growth. This maintenance will resolve disk io performance, occasional power on issues, server hangs and other performance issues.We will begin tonight at 8PM EST and will continue through the weekend.
NOTE: No downtime will be expected.
If you have any questions please let us know.
CloudAtCost Team.
Sounds like we should expect downtime then?
-
@coliver said:
Sounds like we should expect downtime then?
Based on past performance, probably so...
-
@Danp said:
@coliver said:
Sounds like we should expect downtime then?
Based on past performance, probably so...
You mean you don't trust there Note: No Downtime will be expected? haha.
Maybe I'm the only one who likes under promising and saying downtime is likely at any scheduled maintenance time. That way people expect it if it happens. If it doesn't they think it's even better (or just don't care).
-
I agree, warn of possible downtime so that people are happy when nothing blips.
-
Major IOWait issues today. This was during a 1GB Tar operation:
06:50:02 PM all 14.62 0.00 1.91 61.34 0.00 22.13 07:00:02 PM all 4.97 0.00 0.78 30.17 0.00 64.08 -
Major IOWait issues tonight. This was during a 1GB Tar operation.
06:50:02 PM all 14.62 0.00 1.91 61.34 0.00 22.13 07:00:02 PM all 4.97 0.00 0.78 30.17 0.00 64.08 -
Ouch.
-
yeah this started around 4:30 CDT. I was working at 4 and it was fine.. my session started dropping and when it would connect was slow as dirt.
-
My SFTP uploads time out on me most of the time anymore on the one box I use the most.
-
@thecreativeone91 said:
My SFTP uploads time out on me most of the time anymore on the one box I use the most.
Had that happen consistently last night too. Have local network issues so don't want to report that yet. Did a large datacenter (Toronto) to CloudatCost transfer over SCP (basically SFTP) over the weekend without an issue. So think that the issue might be local, but will need to test.
-
@Dashrender said:
yeah this started around 4:30 CDT. I was working at 4 and it was fine.. my session started dropping and when it would connect was slow as dirt.
Ah, well that could be what I was seeing both with the IOWait issues and with my SCP sessions failing.
-
There were definitely issues at C@C yesterday. I haven't logged in yet this morning to see how things are going.
Definitely finding that C@C can't be used for anything more than testing at this point. We've been talking about them for over a month and a week doesn't go by that they don't have problems.
-
I think that you can use it beyond testing / lab use, but only for limited extended use at this point. An obvious (more or less) use case would be a non-real time decision cluster for something like Hadoop where the hiccups are not of critical concern but the cost savings is. You wouldn't use this for processing for trading applications, even Hadoop decision stuff, but if you are doing slightly less critical data analysis using massive Map/Reduce systems, I think that this fits well even for production use.
But, in general and mostly for the SMB, CloudatCost is the ideal lab platform for sure. But they need to mature a bit before they are ready for traditional production workloads. But it really is a beta at this point, so I'm not surprised. At the price, we are going to be deploying on it more and more. I expect we will be at thirty of more servers there by the end of the year.
-
@scottalanmiller I see my server there as a random lab space. I can spin things up and down faster than I can on my internal VM infrastructure because I have never setup templates.
-
@AmanBhogal said:
@thecreativeone91
Not at all about filterning through the PR team. matter of fact, I kinda am the PR team.
It's more of making sure we make the right decisions and make the right steps to move forward.
Two months, another outage.... and still no follow up here.
-
@AmanBhogal said:
@thecreativeone91
I will get you answers! just give me a couple of daysTwo months, that's more than a couple days. Any word on those answers?
-
@AmanBhogal said:
Hey guys!!
I just spoke with the team. We have some hardware upgrades happening this week that will likely increase any IO issues you may be having. in short, we have just ordered a ton of hardware to be able to provide better service and keep everybody happyedit: I selected my words poorly. i said increase IO issues, when in reality I meant that it will help resolve issues
IOWait issues were worse than ever right before the service was shut down. This fix never happened.
-
They're Canadian, right?
Why haven't they tweeted #sorry ?
I thought that was a stereotypical Canadian tagline. Perhaps I'm incorrect about the tag line or their Canadian legitimacy.
