What a bad day. Looks like when they re-imaged my Xeon to install cpanel, they may have killed or erased my slave drive with all of my data.

Has anybody else had this happen?

I had all of my recent backups there. I even stated in my Upgrade ticket to not do anything with the 2nd hard drive. In the Support Chat, the Tech tells me there are no partition tables, and no way for me to get the data,

Is there anything that can be done? Either they erased it, or it failed at the same time as the update, or that is not my drive in the slave position.

Thoughts? Am I screwed? Not very happy at the moment.
Glenn

Gotta hand it to tech support, they sure do try to help.

#$@% happens, and I do appreciate the helpfulness.

Kudos guys/gals.

Hoping tommorrow is a better day.
Glenn

Gosh, did they tell you why they did that? even if it was a non-intended mistake (which of course it was)

How is it that this occurred? and most important... what can be done so that it doesn't happen again?

No idea why/how it happned. I even specified in my Upgrade ticket NOT to touch the slave drive.

It seems the partition table for that drive was deleted. Still working out how to best try to recover the data.

Sooo, not much fun in my book.

went for a Priority 2 hour restore. almost 4 hours later, I get the machine back restored.....to Ensim.

How does that happen? Seems the accounts do not get updated quickly enough. My account still said an Ensim server.

Just a very bad day all the way around. Now I get to wait for yet another restore so that I can start rebuilding my accounts.

Thanks for listening to me vent. Between the Cpanel upgrade fee, data erased in the process, a priority restore fee... and now more waiting for the correct restore..... arrgghh An expensive day.

Vegas sounds good right now... lol

SB

Vegas would be a disaster for you. There's always tomorrow

SB

Vegas would be a disaster for you today, there's always tomorrow

oops!

Sir,

While it's regrettable there's no data on your second drive, I personally performed the first restore and I am 100% positive your second drive was not touched. I remember the specific notes in your ticket not to touch the second drive. What did you have the second drive mounted as? Is it possible it was /home2 and the drive wasnt properly mapped to it?

Hi Patrick,
Thanks for the response.

Yes, it was called home2 before the re-image. Not sure what you mean by maybe not being properly mapped, however it did work how it was. Nightly backups where written there, previous cPanel backups where there from when I moved to RS.
Glenn

We certainly regret any loss of data. Offsite backups in such an upgrade are critical.

We are somewhat new to Cpanel and just learning many things. However, in this instance, you had Patrick on the upgrade and he specifically left your second drive alone.

In our system, we do not change the "type" of server that you have, only add Cpanel as an added service. We'll need to look at how we handle the restores to make sure we are restoring the proper OS.

In the next few days or week or so, we will have a backup service available where you can point yoru backups to our ftp server that is in front of a Blue Arc massive 10 terabyte disc array.

I do apologise that you lost any data we we would not want you or anyone to have that happen. We have been through it personally and know that it is no fun.

I'm planning on ordering the upgrade and also have a second drive on my P4.

I was hoping to move to RH 7.3. Is this possible? Also, how long will the server be down?

Thanks for the post Headurfer.

I do have most everything backed off site, just not as recent as I would like.

I don't doubt Patricks abilities. "Something" went wrong. I don't want to point fingers, and that is not my intention. Just a little bothered that my plans to make this smooth are now toast.

The support guys Josh, Jeffery and Joe have been great in helping last night. Good group of people. While they could not fix it at the time, they put in a lot of effort on my behalf.
Glenn

HS: One note about your backup system. As far as I know, CPanel can only backup locally to a hard drive in the system, not to an FTP server as Ensim does. This would make autobackups with CPanel to the blue arc 'harder' for the newer user (yes I know you can just write a bash script... but most people dont know how). Any comments/ideas?

I guess we should have taken the time to have sucked the rest of your sites up to my box. I guess hindsite is always 20-20.

What I mean is if the directory was still just a directory on the first drive and not mounted to the secondary on accident, and therefore the second drive wasnt really being used. Just a thought, in reality we have no idea.

Now I understand what you mean.

Pretty sure it was actually being used. It showed disc space being used in the expected amounts....but it does not much matter at this point.
Glenn

OK, after much trying, reading (google is my best friend) I tried a different method to "guess" at recreating the lost partition table.

Wow, I was finally able to mount the thing (was very surprised when no errors came up when I gave the mount command..LOL). and
There is my data!! I could see it all, and got pretty excited. I was even able to transfer some of it to the main drive. So, I am a happy guy. A lot of the data wont move, but for now... it is better than nothing.

Incase anybody cares, the data that won't transfer gives this error:
cp: cannot stat `backup/2003-03-20': Input/output error
cp: cannot stat `backup/2003-03-22': Input/output error
cp: cannot stat `backup/2003-03-24': Input/output error
cp: cannot stat `backup/2003-03-25': Input/output error
cp: cannot stat `backup/2003-03-28': Input/output error
cp: cannot stat `backup/2003-03-29': Input/output error
cp: cannot stat `backup/2003-03-30': Input/output error
cp: cann

Congrats! I am glad you have been able to get your hands on the `lost' data.

For the benefit of all of us here, can you pls post the link or the method that you used to recover from a lost partition table?

Sure Of course, use at your own risk and all of that...

http://www.lissot.net/partition/partition-6.html

wasn't perfect, but was enough.
Glenn

I have almost exactly the same experience.

Went for priority 2 hour restore, restore happened 10 hours later. I had a CPanel server, restored to Plain Redhat. I had a slave backup drive which I specifically put in the ticker not to touch, and it was completely missing after CPanel was put back in.

I am still waiting to find out the status of the backup drive. I haven't slept for a long time staying up to reset my server, which can't happen without the backup. I was told by overnight support that data in the backup will be safe if I put it in the ticker.

We shall see what happens...

Erwin, I feel your pain.

Hopefully, your drive is intact, and can be mounted as /backup
If the partition tables where overwritten like mine, then you are in for some fun (not). I was lucky, and recovered some critical data from my drive, but it was a pia.

Still, even with the *problems* I encountered with the upgrade, the support crew was very helpful. Some when above and beyond to help, and it is appreciated. Jfeffery and Josh in particular.
Glenn

LOL! My sorry saga continues.

They had the hard drive in there but said they didn't reconnect.

So they did.

In the meantime, I caught an hour of sleep.

I was pleased to hear the hard drive was reconnected. So I tried to SSH in - no host found. My CPanel, WHM and SSH didn't work.

So I had to open another ticket for a reboot.

After some waiting, the reboot happened.

But the 2nd drive is still missing.

Strange but true. So my site is now offline close to 20 hours now... and I'm still back at square one. At least CPanel works.

Erwin, bummer.

Did you try to mount the second drive? If so... and errors?

I can't mount it because it doesn't exist.

According the WHM server info, there's only one drive.

When I do a "df" in SSH, there's only the main drive.

it won't show up in WHM if it is not mounted. This may work:

cd /
mkdir backup
mount /dev/hdc1 backup
cd /backup
ls

mount: /dev/hdc1: unknown device

Like I said, it looks disconnected or something.

Even if it doesn't show in WHM, it should come up in the mount/ format section anyway, shouldn't it?

And it should come up using "df" in SSH.

What happens if you do this:

fdisk /dev/hdc (hit enter)

then: p (show partition info)

(this assumes your system is using your second HDD as hdc)

Glenn

I get this:

Unable to open /dev/hdc

I seriously think it's gone. The response from the ticket is:

"sending to webtech for investigation."

Whatever that means...

Meanwhile, here I am 20 hours in front of my computer, wanting very hard to put my site back online...

Erwin, I hear ya.... I wish I could help more.

Might jump in the IRC chat, see if Josh can have a look when he is not too swampped. He is there now.
Glenn

I'll try...

The support people in there are run off their feet. I don't think Josh has time to help. I've PMed him in there anyway.

This is depressing - my site hasn't gone this long offline ever.

Erwin, I know.... have been watching the room. Quite busy today.

Josh should be able to point you in the right direction... he is very helpful.... but swampped it seems.
Glenn

He's fixed it.

I'm off to set up my site...

Awesome news! Glad to hear it!

Glenn

Thanks SB-Host
I was able to mount 2nd drive without format.

So why not just ask support to unplug the drive itself (power wise) and that way you have a 100% guarantee it will not be touched on a restore.

Then you have to remount it because the system will not know its there.

Regards.

And why do things like this take 20 hours to fix? If I went 20 hours, I would lose half my clients!

Well, things just went wrong, in series I guess. I was in communication with Rackshack during the problem, and was in contact with my customers thoughout the whole time period. I did not lose anycustomers in the end. Communication was key in my opinion.

Yeah, it sucked. It was no fun. Rackshack got it right in the end, and a few Techs went out of their way to help. I was able to recover the data off of the second drive. As a result, I think Rackshack looked at how they handle the upgrades, and probably put in some good effort to ensure it does not happen again.
Glenn