Spectre and Meltdown - The what is my provider going to do about it? thread!

WSS

This is fun.

$ uptime && grep -E 'cpu_insecure|model\ name' /proc/cpuinfo | sort | uniq
 16:05:00 up 38 min,  0 users,  load average: 0.56, 0.65, 0.69
bugs        : cpu_insecure
model name  : Intel(R) Core(TM) i5-2540M CPU @ 2.60GHz

I'm at .5 load doing nothing on likely the thinnest distribution possible with functional X11.

E: Wrong thread. Whoops.

joepie91

@Neoon said:

@Clouvider said:

Neoon said: How many containers do you need on a single node that you need 60min+ to reboot it?

1 Windows KVM that refuses to budge. And then you have a choice. Downtime or potential loss of data?

Well I said containers, most of them where OVZ boxes.

FYI, I've seen entire host nodes being kept from shutting down because of a single OpenVZ container that refused to die. This issue absolutely exists with OpenVZ as well.

doghouch

@joepie91 said:

@Neoon said:

@Clouvider said:

Neoon said: How many containers do you need on a single node that you need 60min+ to reboot it?

1 Windows KVM that refuses to budge. And then you have a choice. Downtime or potential loss of data?

Well I said containers, most of them where OVZ boxes.

FYI, I've seen entire host nodes being kept from shutting down because of a single OpenVZ container that refused to die. This issue absolutely exists with OpenVZ as well.

Stuck? Can’t you kill their init process and get it over with?

WSS

@doghouch said:

@joepie91 said:

@Neoon said:

@Clouvider said:

Neoon said: How many containers do you need on a single node that you need 60min+ to reboot it?

1 Windows KVM that refuses to budge. And then you have a choice. Downtime or potential loss of data?

Well I said containers, most of them where OVZ boxes.

FYI, I've seen entire host nodes being kept from shutting down because of a single OpenVZ container that refused to die. This issue absolutely exists with OpenVZ as well.

Stuck? Can’t you kill their init process and get it over with?

I think you missed the entire bit about having a clean shutdown.

doghouch

@WSS My mistake: screw VZ.

WSS

@doghouch said:
@WSS My mistake: screw VZ.

I joked about just doing a shutdown -r now, and letting fsck take care of it if you're still using simfs. That'd probably be one of the fastest ways to get things back up.. unless you know, you're in Romania.

doghouch

@WSS said:

@doghouch said:
@WSS My mistake: screw VZ.

I joked about just doing a shutdown -r now, and letting fsck take care of it if you're still using simfs. That'd probably be one of the fastest ways to get things back up.. unless you know, you're in Romania.

What? fsck only took a whole goddamn month.

(on another note, fsck_hfs was a godsend when Windows went retard mode on my MacBook.)

IAlwaysBeCoding

@Neoon said:

@WSS said:
Oh, look, it's time for some @Neoon rage. Which project are you going to abandon now?

@Neoon, you have used that gif way too many times here.You are one of the members from LET that recycles the same gif memes over and over again, it's get old very fast. Why can't you just construct a proper reply, instead of relying on same old boring ass memes? Please don't give me that bullshit response that " an image is worth 1000 words" reply.

@WSS is absolutely right, everytime you have one of those rage fits you end-up taking it on your projects and just throw them away. Heck, even @MasonR is planning on building an open-source project for node uptime checking, and you don't even want to open-source your shit so the community can benefit from.

Neoon

@IAlwaysBeCoding said:
@Neoon, you have used that gif way too many times here.You are one of the members from LET that recycles the same gif memes over and over again, it's get old very fast. Why can't you just construct a proper reply, instead of relying on same old boring ass memes? Please don't give me that bullshit response that " an image is worth 1000 words" reply.

IAlwaysBeCoding

@Neoon said:

@IAlwaysBeCoding said:
@Neoon, you have used that gif way too many times here.You are one of the members from LET that recycles the same gif memes over and over again, it's get old very fast. Why can't you just construct a proper reply, instead of relying on same old boring ass memes? Please don't give me that bullshit response that " an image is worth 1000 words" reply.

Oh look! is another meme image from @Neoon. Why am I not surprised?

Neoon

@IAlwaysBeCoding said:
Oh look! is another meme image from @Neoon. Why am I not surprised?

LosPollosHermanos

@ramnet said:

Backups nodes is overkill when it takes literally 10 minutes to apply the patch and reboot.

Your nodes must not have many VPS's on them or maybe you are using SSD RAID or something. I have some nodes that take 30 minutes or more to complete a reboot cycle.

jackb

@doghouch said:

@joepie91 said:

@Neoon said:

@Clouvider said:

Neoon said: How many containers do you need on a single node that you need 60min+ to reboot it?

1 Windows KVM that refuses to budge. And then you have a choice. Downtime or potential loss of data?

Well I said containers, most of them where OVZ boxes.

FYI, I've seen entire host nodes being kept from shutting down because of a single OpenVZ container that refused to die. This issue absolutely exists with OpenVZ as well.

Stuck? Can’t you kill their init process and get it over with?

Not always. Once I saw a vps that its init was gone but a child process was in uninteruptable sleep and refused to be killed. Iirc something to do with networked filesystems and fuse. Nothing brought it down until the next HN reboot.

LosPollosHermanos

@Neoon said:

@WSS said:
Oh, look, it's time for some @Neoon rage. Which project are you going to abandon now?

I am outraged at Intel and AMD, and ARM manufacturers. Time to use something else! Also the internet. I am outraged that computers that are accessible by almost every human on the planet will always have
some security risk. That's unacceptable! Damn you internet.

tarasis

Mmm I've only heard from one provider so far.

VPSDime - got email from them on Saturday to tell me that they would be rebooting nodes on Saturday to apply an update, and to see https://vpsdime.com/announcements.php?id=158 for any future information.

Nothing from

• Online.net (have a dedicated there)
• VirMach (OpenVZ and KVM)
• HostHatch (KVM)
• XVMLabs (although the VPS was rebooted without warning yesterday)
• Time4VPS (OpenVZ)
• Aruba (VMWare)

Shazan

@tarasis said:
Nothing from

• Online.net (have a dedicated there)

Being a dedicated server, it is up to you to update it or not. I don't think they will do anything about it.

JoeMerit

@Virmach can I expect a reboot soon on my KVMs ?

Francisco

We'll be starting this week, I was mostly wanting to wait to see if there was any big issues that came up from everyone else moving over.

There seems to be some XEN conflicts, but since we don't XEN anywhere, it should be OK.

Francisco

rick2610

@Shazan said:

@tarasis said:
Nothing from

• Online.net (have a dedicated there)

Being a dedicated server, it is up to you to update it or not. I don't think they will do anything about it.

online.net has a detailed status page: https://status.online.net/index.php?do=details&task_id=1116

Hxxx

Don't worry about it guys, providers don't need plebs to tell them when or how to apply patches. Just stating that, I mean only applies if you are with a real provider and not some summer host.

Hxxx

Nice meme sensei.

@Neoon said:

@IAlwaysBeCoding said:
Oh look! is another meme image from @Neoon. Why am I not surprised?

tarasis

@Shazan said:

@tarasis said:
Nothing from

• Online.net (have a dedicated there)

Being a dedicated server, it is up to you to update it or not. I don't think they will do anything about it.

I wasn't expecting them to do anything, other than emailing me (mass mail) to inform me/us that I should aim to update to a kernel that includes the fix. I'm sure there are some Online.net users that aren't aware of the issue.

@rick2610 said:

>

online.net has a detailed status page: https://status.online.net/index.php?do=details&task_id=1116

Cheers!

Maounique

Yes, Xen we leave last too. we havent decided yet how to deal with this and our older pv issue.

mksh

@huntercop said:
sudo (Holy shit that's half of it)

Fuck sudo.

apt-get (common sense part) update (omg that simple?)

Yeah, that simple but you still haven't upgraded your system.

joepie91

@jackb said:

@doghouch said:

@joepie91 said:

@Neoon said:

@Clouvider said:

Neoon said: How many containers do you need on a single node that you need 60min+ to reboot it?

1 Windows KVM that refuses to budge. And then you have a choice. Downtime or potential loss of data?

Well I said containers, most of them where OVZ boxes.

FYI, I've seen entire host nodes being kept from shutting down because of a single OpenVZ container that refused to die. This issue absolutely exists with OpenVZ as well.

Stuck? Can’t you kill their init process and get it over with?

Not always. Once I saw a vps that its init was gone but a child process was in uninteruptable sleep and refused to be killed. Iirc something to do with networked filesystems and fuse. Nothing brought it down until the next HN reboot.

Yep, that's exactly what I've seen as well, although I don't think it was FUSE that time. No chance of a clean reboot when that happens...

Zerpy

@tarasis said:
I wasn't expecting them to do anything, other than emailing me (mass mail) to inform me/us that I should aim to update to a kernel that includes the fix. I'm sure there are some Online.net users that aren't aware of the issue.

They actually will email people when there's patches available for their servers for all 3 vulnerabilities - they could send an email already saying: "There might be a kernel update available, but no microcodes, so patching now would be useless

Their blog:

We will send an email to all our Dedibox customers when we will get all the microcodes and the updated kernels.

WSS

NetCup just restarted one that I forgot to last night before bed, and both Inception and ExtraVM have posted their time-down frame. Overall, I'm not displeased, but I do question a couple EU KVMs I have which have been completely silent on the matter. I don't think Dedicenter (lol) is even aware that their computers are computers.

hostdare

WSS said: I don't think Dedicenter (lol) is even aware that their computers are computers.

Your favorite host

WSS

@hostdare said:

WSS said: I don't think Dedicenter (lol) is even aware that their computers are computers.

Your favorite host

If they weren't so unabasedly "no shits given", I would give them some slack.. but they just don't give a flying fuck.

Radi

@Francisco said:
There seems to be some XEN conflicts, but since we don't XEN anywhere, it should be OK.

I would love to hear about that as most of our deployment is Xen-based. Can you PM me with some details?