[rossdev]

This was an interesting read... http://yoast.com/vps-net-issues/

[Steve]

Was Yoast really a client "from their very beginning in fact"? I don't think so.

[Terry Myers]

Steve, on 10 September 2011 - 09:26 AM, said:

Was Yoast really a client "from their very beginning in fact"? I don't think so.

June 18th, 2009. He's pretty close....

[boeki]

a guy commented there that he hasn't had a downtime in the past 12 months.

i asked what cloud he's using so i can move my VMs there. still waiting for a reply.

i also twitted vpsnet a few days ago asking them to recommend a stable cloud. still waiting for a reply.

my VMs are on slc-d and atl-e.

[Nick]

Being a customer for about 1 year and 9 months, I have had both really good and bad times. A while back I have had around 100 days uptime on my machine in LON-C, which is really good I think . But after that I had to reboot it for updates, and never even came close to great uptime.

The last few months have been really horrible after i made the move to AMS a while back. At first the speed was really good. The I/O was awesome, and load times dramatically improved.

But then problems started . Sometimes HV's just crashed, and my vps was either restarted, or it was stuck because it wouldn't start, and it needed an FSCK. Then really nasty problems occurred, my vps went into read-only mode lots of times. As my websites aren't that critical I just monitor for ping, checking if the machine is alive. But the problem with that is, that in read-only mode the vps still responds to ping's so it sees the machine as being up. It still ran, but just could read from the SAN, because it lost the connection somehow, or the SAN had a problem. So basically in read-only it turned into a zombie, still running, but serving no websites at al.

This happened to me multiple times (10-20 in about 3 months), causing me hours of downtime up to half a day, until I noticed it . Most of the times when I was asleep. On top of that sometimes one of the two redundant SAN's acted up. Then the whole process of synchronizing these SAN's for data redundancy had to start, causing painful poor performance, and almost a locked up vps, again hardly serving any websites.

I was just about to move until I got the news on the storage platform being replaced, and the network being upgraded. Still seeing the good in vps.net, as I don't always blame them for the soft- and hardware failures, I gave them another go. I guess it's a learning experience for everyone, things go never as you planned or hoped.

But maybe It's a little early to say, as updates on AMS just completed. But it looks like the stability has returned once again for me. The load has gone down a lot, close to 0, as I don't run intensive processes, and the I/O is a lot better. I'm seeing the %iowait once again very low at about 0.39%, which is very good I guess. For now we'll see what happens, and hopefully this stays. Hopefully I can report back in a while, still seeing improved result, and being a happy customer.

Any of you guys have had similar experiences in the past, and seeing improved results now ?

P.S. Hopefully I didn't jinx myself now

[Miikka]

Nick, on 11 September 2011 - 09:45 AM, said:

As my websites aren't that critical I just monitor for ping, checking if the machine is alive. But the problem with that is, that in read-only mode the vps still responds to ping's so it sees the machine as being up. It still ran, but just could read from the SAN, because it lost the connection somehow, or the SAN had a problem. So basically in read-only it turned into a zombie, still running, but serving no websites at al.

And if you monitor that web server is serving content then it will continue still while filesystem is read-only - only if you have writes happening in your web service then it would fail. Naturally at the same time you would be losing all log content from web server as it cannot write to the filesystem.

I solved the filesystem monitoring with Server Density. I created a process which was writing to a file every minute. If it failed, the process terminated itself immediately. At the same time, Server Density monitored that this process was running and alerted me immediately if it detected that the process wasn't running.

This post has been edited by Miikka: 11 September 2011 - 11:23 AM

[Nick]

Looks like the stability has returned to AMS, everything is still running very smooth. Google is reporting my page load times at 1s now, 90% faster then other sites. It was a lot higher the past couple of weeks.

Anyone seeing similar results in AMS or other improved clouds?

[Spiral]

Interesting thread and I believe this topic is now today even more relevant now than when
this discussion was first opened back in September of this year which at the time we were
about at the end of the first month of the present problems and issues ....

Steve, on 10 September 2011 - 09:26 AM, said:

Was Yoast really a client "from their very beginning in fact"? I don't think so.

I don't know about Yoast but we definitely have been with VPS.NET from the very beginning here!

I have several accounts active now and several accounts before those and our original first
account with VPS.NET was back about a little more than a full year **before** the memory on
the nodes was changed to 376 MB which was now quite a while ago back when the nodes all
had 256 MB of memory at the time back then.

In all the time I've been with VPS.NET for the **MOST PART** it has been extremely reliable
and the only time problems really came was when they were trying to convert to new technologies
mainly just that whole SAN 1.0 to SAN 2.0 fiasco that blew everything up everywhere a long
while back which until just recently was actually the most significant problems I ever saw
with VPS.NET. The rest of the time any instance of problems or outages few in number and
spaced very far apart and a very rare occurrence and that is actually one of the very reasons
we've kept such a long standing relationship with VPS.NET with much of our business.

And then comes the past 2 1/2 months since this downward turn and the story continues ....

Utter complete hell for lack of a better word. Constant outages, constant servers offline,
repeated high loads, network dips, and read only file systems. Every single day the past
2 1/2 months I have had server outages which sometimes show up on the status.vps.net page
and many times nothing is posted on the page whatsoever except that I have so many servers
across so many accounts that it is very easy and obvious for me to see first hand which
clouds are actually having trouble so I know first hand how many of the cloud outages
aren't even being posted or reported to the status page at all.

The first month of the current new problems and outages back when this forum topic was opened,
I had a lot of angry customers but most all stuck through those long outages and stayed
loyal and hopeful that the new found problems would soon resolve and I myself relying on
VPS.NET's mostly excellent previous past history over the past several years, I also
hoped the same and have like my customers done everything to give them a fair chance to
get the various problems and issues plaguing the service all resolved completely.

Then last month, the situation was even worse with more outages and broken clouds being
a now daily occurrence with some lasting a full day or more many our own clients gave up
and finally left and personally I don't blame them any. It's embarrassing the number of
outages we have seen around here from the beginning of August until now.

Again I stuck things out but at the same time started looking at alternative possibilities
for all our many, many servers and customer sites as well as all of our own sites too.

And then comes the past couple weeks through October which has been nothing short of a
non-stop roller coaster nightmare and also lost another large wave of clients and this
week's outages are just inexcusable and I am really starting to lose hope here that
VPS.NET is in fact going to get their act together and get all the recent issues sorted
and now even more so more than ever really squeezed into the position of either choosing
giving up on them entirely or give them just a bit more time to sort the problems ....
That, or finally bidding VPS.NET a final farewell in a massive mass exodus style, which
is a fairly apt description in my case considering just exactly how very many servers
I and all my companies directly control around here.

A few days to maybe even a week of random problems might be something but this has been
a plague going on for now more than 2 1/2 months and every time it seems things might
just be finally starting to be back on the up swing again, you get outages like Atlanta A
just recently this very week up through yesterday and right now at this moment Chicago B
(which has one of the longest most stable histories) is showing the initial signs of all
the other clouds that have collapsed one by one over the past several months and the same
thing I was seeing in Atlanta A this past weekend just before it went down and remained
offline all day long so apparently the saga continues.

Well there is my 2 cents on recent network troubles and I continue to hope that things
will soon be better around here though I like probably most others here cannot keep on
waiting forever and ever endlessly.

If there is hope that VPS.NET will sort things out and get things back running as well
again as they once were before all the present troubles began, I'd be willing to stay here
and give them the fair chance they have earned over the years but also at the same time
I like most here also have to weigh all the damage to own business interests and ultimately
it's going to be purely a business decision what happens next verses any personal
feelings or sentiment.

Please fix the networks and do it soon!

This post has been edited by Spiral: 19 October 2011 - 12:13 AM