VPS down all day and being unreliable

[srichards]

My vps has been off for most of the day. Email often inaccessible. Plesk not responding, websites not responding. For the last two days I have had to restart qmail or imap to get any sense out of email.

Support claims everything is fine. I have had one ticket response 6 hours (!) after the initial complaint. What happened to the 2 hour response time???? Ticket response gives me some gibberish about updating the kernel and that is why it was only showing 3GB free. Totally unrelated to my complaint. I said email and web and plesk were off. I mentioned Load Average being low and asked whether someone else's vps was the root of it as my vps was like a snail with heavy shopping. Memory looked fine. I've only got a few wordpress sites running on it. It hasn't been reliable enough to put anything important on there. It briefly returned to life around lunchtime but has gone for a nap since then.

It's not even responding to a ping at the moment.

[eUK-Alan]

My vps has been off for most of the day. Email often inaccessible. Plesk not responding, websites not responding. For the last two days I have had to restart qmail or imap to get any sense out of email.

Your vps and all other vps were working fine after you reported about downtime of your vps . The main reason of downtime for your vps was incorrect memory shown by server because of old virtuozzo kernel. Due to this, the server was rebooting all of sudden at regular intervals.

Support claims everything is fine. I have had one ticket response 6 hours (!) after the initial complaint. What happened to the 2 hour response time???? Ticket response gives me some gibberish about updating the kernel and that is why it was only showing 3GB free.

We have been working on server to fix this issue since you and a few other clients reported us on live chat / email. (We did not update your support request as your vps was running until we started kernel upgrade)

Just to keep you updated we are still working on this issue. The server rebooted few times but after reboot its back on old kernel only and not on new kernel. The problem hasn't been solved yet but we will fix it shortly.

We will keep you informed about the progress in ticket.

[eUK-Ken]

Hello,

The main server i.e the main hardware node is still facing problem i.e booting on the Main Virtuozzo Kernel. So now we have decided to reload the server and all the VPS's will be restored from the latest back i.e of today few hours before this mishapp.

Regards,
Ken,
Support Team.

[srichards]

Any idea when this will be completed as downtime is now over 12 hours?

[Jester]

Just for reference, I am in the same boat - with over 12 hours downtime.

Customer Support has been excellent though..

[eUK-Ken]

hello,

Yes James Your VPS shall be up and running in the next 40 to 50 mins. Restoration is in process.

Regards,
Ken.

[eUK-Ken]

Hello,

We are in the process of restoration of each VPS one by one. So all the VPS will be up and running in the next 4 to 5 hours.

Regards,
Ken.

[Philip S]

Hello,

We have successfully restored and restarted all the VPS on Main Hardware Node.
We have just finished doing so and everything should be completely stable again now.

[srichards]

Unfortunately my vps is unreachable again - ignoring pings totally as of about 22:40. Email disappeared first which is how I noticed... still dead at 22:55.

[srichards]

I take it back. It's Virgin Media's shoddy network routing again! I'm cut off from my own server just like has happened to other people. I've got a monitoring service it and it is showing it as being up now. Can't ping it from VM though. GRRRRRR!

[Jester]

Really? Because my server appears to have fallen over again... I'll have a look to see if it is a Virgin problem..

[srichards]

It seems to be a bit of both. I've got monitoring on ippatrol now to cross check what is going on. Server went off somewhere between 22:30 and 22:40 and so did my connection to it from my broadband. Server was ok not long after but broadband connection to it only returned at about 23:20.

I'm wondering whether there are some routing algorithms around that are preventing connections to servers that have recently been off which would explain the disparity.

I shall check the log files as that will say when it was up and down.

Are you with virgin media?

[srichards]

reboot system boot 2.6.18-028stab06 Mon Jul 19 18:19 (00:22)
reboot system boot 2.6.18-028stab06 Sun Jul 18 23:04 (19:36)
reboot system boot 2.6.9-023stab048 Sun Jul 18 08:14 (1+10:27)
reboot system boot 2.6.9-023stab048 Sun Jul 18 04:08 (1+14:32)
reboot system boot 2.6.9-023stab048 Thu Jul 15 21:28 (2+06:36)
reboot system boot 2.6.9-023stab048 Thu Jul 15 20:38 (2+07:25)

All times are EDT for some reason. I set it to GMT/BST but it seems to have been changed.

Hmm. Not really sure what to think now as logs are showing varying states of whether it was up and down compared to my own checks. From the ippatrol it seemed to show that it was up as far as some routes were concerned but not others so I have no idea whether it was up or down!

[Jester]

I'm with Virgin Media.


Spoke to live chat and they said a senior VPS admin was updating the kernel or some-thing-or-other.


Server is back up now. Better to do maintenance now, rather than in office hours...

[srichards]

True. But it would be better if they mentioned they were doing maintenance, especially when a customer has experienced down time as it makes them wonder whether the service has gone to pot!

I hope this isn't the same kernel upgrade that failed before.....

[eUK-Ken]

Hello,

I do apologize for this inconvenience caused to you. We had forwarded this issue to SWsoft engineers and they were working on this kernel upgrade and have rebooted the server for the emergency purpose.


Regards,
Ken.

[srichards]

It's just gone off again! Was trying to send an important email and bam. Disappeared. The irony was I was also about to write to someone in parallels about my experience with plesk.....it's going to be even less positive now!

No pings being answered so it's properly off.

[Jester]

Yep, same mate - apparantly the node is being rebooted again. It went down last night too at about 1am - I was told only my services were rebooted.

There is something seriously wrong with this node. Over 12 hours of downtime was bad enough last sunday.. having to ask livechat to reset the server every few hours is getting beyond a joke. It makes me look like a joke to my customers.

Something needs sorting, and it needs sorting now.

[Rock]

It's just gone off again! Was trying to send an important email and bam. Disappeared. The irony was I was also about to write to someone in parallels about my experience with plesk.....it's going to be even less positive now!

No pings being answered so it's properly off.

Yep, same mate - apparantly the node is being rebooted again. It went down last night too at about 1am - I was told only my services were rebooted.

There is something seriously wrong with this node. Over 12 hours of downtime was bad enough last sunday.. having to ask livechat to reset the server every few hours is getting beyond a joke. It makes me look like a joke to my customers.

Something needs sorting, and it needs sorting now.

Hi,

The problem with the host server is taken care of now. I've asked one of the senior VPS administrators to look into this on top priority.

[srichards]

Cleary this node is sick. All those affected should be moved to a new node asap with their current ips retained.

The last but one kernel version used was much more reliable. Perhaps going back to that one might be better??? I even wonder if parallels test their new kernels on *live* machines that are in proper use by normal people. They might work perfectly on a fresh installation that hasn't been altered from default settings but no one keeps their settings exactly as they are out of the box!