Don't know what is happening

[tiserves]

I have been one of the biggest proponents of this service since I signed up a few months ago, but a couple events in the past 24 hours lead me to believe that a couple things could be done to improve the service even more.

Yesterday I received an e-mail stating that there would be "scheduled maintenance" on my server. I got only a couple hours of notice about this, not even a day's notice. I went to work on the server later and found out that the planned maintenance had not occurred some hours earlier, since I had 13 days of uptime. So I started working on my sites. In the middle of this, I received notice that the server was going down immediately. I quickly exited and waited. The service was supposed to take about 20 minutes. I waited for over an hour, then figured that as things do, something was taking longer than expected.

Stupidly, I forgot about it after this, and didn't think to check on it until this morning.

Almost 24 hours after what was supposed to be scheduled maintenance, my sites are down and my Teknic panel is unreachable. Does Unixshell have any way to determine whether or not the individual sites are up and running after you perform maintenance?

I tried to post a question on the "outage" section asking if anyone else was down, and I got a response that this section of the forum is now MODERATED. I understand the need to have a section for admin-only topics regarding outages, but perhaps underneath that forum you could have another for us lowly users?

I would just like to make a couple suggestions... First, when you're performing what appears to be elective maintenance (I believe this was for a kernel upgrade?), please give us at least a day's notice. Secondly, I hope you can find a way to make sure all virtual machines are back up and running when maintenance is finished.

You mentioned one of the big benefits of this "daytime maintenance" would be staff that's alert instead of tired after a day's work.

Apparently, someone was still asleep at the switch.

[tiserves]

I just got a response back from Support:

"Your vps was up but required a fsck."

I don't know what good it is to have the server up if I can't access it. Is it my responsibility to perform this fsck?

I will say one thing that's still valid with Unixshell -- support tickets are answered very quickly.

[matta]

Yes, we provide all the tools to access and manage your server. Our staff will do tasks to bring your VM backup if a ticket is submitted, but for the most part at long as our tools (Teknic/SSH console) are working then our hands are clean. To do the fsck all you'd need to do is login to the SSH console, perform the fsck, and then reboot.

[anonet]

I agree that more notice should be given when maintence is to be done (a weeks notice about the upgrade to Xen 3.0.2 would have been nice). On top of that, it would be nice if you have to reboot a box to send a wall to all running VPSs. I realize this is a little extra work but it is a lot better than having customers ticked off because they were in the middle of editing something and it didn't get saved.

As for the "should we have to perform fscks" ... this is an unmanaged service. Of all the VPS providers out there these guys give the most flexibilty when it comes to maintaining your VPS. The cosole shell access is *REALLY* nice. Bottom line, since this is unmanaged it is your resposibility to make sure your VPS can reboot. Or if it doesn't come back up to ssh into the console and see why.

I had upgraded the baselayout in my Gentoo install and screwed up adding the host route in the new net conf file. So when it came time for them to upgrade my VM yesterday I couldn't reach my VPS when it came back up. If this had been ANY other provider, I would have had to open a ticket and have them go in and find out why it couldn't reach the network. That would have cost BUCKS.

In conclusion I just want to say I think these guys offer a GREAT service (just wish they would release the Teknic control panel under an OSS license :).

--Brian

[jcinacio]

[QUOTE=anonet]I agree that more notice should be given when maintence is to be done (a weeks notice about the upgrade to Xen 3.0.2 would have been nice). On top of that, it would be nice if you have to reboot a box to send a wall to all running VPSs. I realize this is a little extra work but it is a lot better than having customers ticked off because they were in the middle of editing something and it didn't get saved.[/QUOTE]

- signed.


In conclusion I just want to say I think these guys offer a GREAT service (just wish they would release the Teknic control panel under an OSS license :).


- signed again :)

[matta]

In Teknic 3.0 (soon people, soon) we will let users keep an e-mail address on file. We will be able to send out a mass e-mail to everyone who has one on file before a reboot so everyone knows.

[tiserves]

[QUOTE=matta]Yes, we provide all the tools to access and manage your server. Our staff will do tasks to bring your VM backup if a ticket is submitted, but for the most part at long as our tools (Teknic/SSH console) are working then our hands are clean. To do the fsck all you'd need to do is login to the SSH console, perform the fsck, and then reboot.[/QUOTE]

Matt, that's just it. I could not even get to the Teknic panel. The ssh address was not responding. When I tried to ping it I was getting "Destination Host Unreachable." If you're telling me that even if my filesystem was shot, I should still have had access to my Teknic panel, then support is not being straight with me as to what was wrong with my account. I was unable to get to the panel, unable to ssh, and therefore I was unable to restart my server in single user mode & do an fsck, which is what I assume you're telling me I should've done.

In short, I had ZERO access to my server for many hours, and there wasn't a thing I could do about it.

Oh yeah, Matt... My server is down right now as well. No access to Teknic panel. 4/24, 6:00PM EST. Can't ssh either (just tried at 6:07pm) Wonder what support will say this time? My fault again?

I'm sorry to sound so negative. I know you guys are working really hard. It's just very very frustrating.

[matta]

I assume someone fixed you up since it seems up now. The host you are on was subject to the random reboot bug posted in the outages forum. I had to reboot it a second time as the correct (updated) kernel was not booted. While we perform a shutdown of the domains we stop the web server on the host and that is why I would have said Teknic was down for you, but 4 hours does not make sense. We used to have a problem with routes not being added properly, but that problem hasn't surfaced for months.

(FYI: You can always access https://vmXX.unixshell.com if you know your host node.)

[tiserves]

Yes Matt, as always, support responded to my ticket yesterday very quickly, so I'm up now. Thank you.

As for the other issue, I don't know what to tell you. All I know is that when your maintenance was over, I hadn't thought to check to make sure my VM was back up until the following day. At that point, I had zero access to my machine in any form. I will keep track of which host my VM is on to use the URL you listed in the future.

Again, I know what I got myself into here, I know where my support ends and yours begins, and I accept that 100%. I've been using Linux and FreeBSD for about eight years now on my desktops and servers. But when it seems like my hosting company is rebooting for elective upgrades almost on a whim with very short notice, claims that it is doing so during the day to make sure its employees are alert, and then doesn't follow up to make sure my machine is up when he (or she) is done, I get a little irritated.

That said, I would still recommend you guys 100% to anyone looking for VM hosting. I'm just trying to point out that there are still areas that need a little tidying up as far as customer support issues go.

Thanks again for listening.

[tiserves]

Just as an appended note to this thread, my server now has an uptime of 20 days. I just wanted to say thanks, and not be one of those folks who complains when things go wrong, but never says "good job" when things are fixed. :)

[jwr]

[QUOTE=tiserves]Just as an appended note to this thread, my server now has an uptime of 20 days. I just wanted to say thanks, and not be one of those folks who complains when things go wrong, but never says "good job" when things are fixed. :)[/QUOTE]

Same here -- uptime of 29 days, which is good. It seems at least some of the stability problems are gone. I'm just slightly worried about the complete silence from unixshell, as well as the message on the front page about the lack of datacenter space (please! does anybody believe that?). I hope they are doing well, I've come to really appreciate the service as it is now.

--J.