Results 1 to 15 of 15
-
12-11-2006, 07:00 AM #1WHT Addict
- Join Date
- Jun 2005
- Posts
- 127
Please advice and help this is happening now and its not good :(
On Sat Dec 09 2006 i asked my provider (layeredtech) this:
"Hi, last night my server went into 'read only file system' this could mean our hard drive is failing, could you please check the hard drive with fsck from the console i wouldn't want the hard drive to fail. thanks."
They promptly answer saying they noticed some I/O errors and offer a solution which was cloning the hard drive with errors into a new one. I accepted.
Then i received this email from one of support:
"You server is back up and running with the old drive. I tried three different cloning options and all of them reported too many bad sectors to be able to clone the drive. At this point the best thing to do is to request a reload and data save and reference this ticket number to have the reload fee waved. Rodney Howard"
Ther server was running now as always but now i knew we have a faulty hard drive so i accepted once again to do the reload expecting also to have all the accounts move but i just received this:
"The accounts will not be on the server. You will need to mount the secondary drive (the old primary drive) and transfer your data onto the new primary drive. If you need assistance with the data transfer I can escalate the issue to our level 3 support, but there will be a fee.
Thank you, Shawn"
Is this fair? it wasn't my fault the hard drive failed also i asked for a solution to the problem not for another problem my server has 2 hard drives i asked a lot of times which one had the problem they never answered i know layeredtech people read this so please help me with this i been a customer for over 2 years now.
-
12-11-2006, 07:05 AM #2The least among you.
- Join Date
- Apr 2005
- Location
- Jacksonville, FL
- Posts
- 981
As much as I hate to disagree with customers, I'm going to side with LayeredTech here. When using an unmanaged host, it's your job to maintain backups and the requisite level of knowledge for systems administration activities in order to manage your own machine. If moving accounts from drive to drive is too much to bear, perhaps you might consider hiring a sysadmin.
Now selling BigVPS's!
Jacksonville Colocation and dedicated servers by colo4jax
We are *not* a reseller. We own our servers, switches, routers and racks.
-
12-11-2006, 07:07 AM #3WHT Addict
- Join Date
- Nov 2006
- Posts
- 126
Originally Posted by otamendi
-
12-11-2006, 07:08 AM #4WHT Addict
- Join Date
- Jun 2005
- Posts
- 127
Originally Posted by tical
They could tell me three of them are online right now, the new one with the OS reloaded, the second one with some accounts (that i had) and the third one with the possibly corrupted data, then i had something to work with when calling our sysadmin to do that, but simply i dont have more info on this... and i asked.
I hope they have a solution and if not, if the data is gone well guess i need another hosting company but i dont want to go before this everything was fine i expect someone from that company reads this, tries to understand me and help.
-
12-11-2006, 07:12 AM #5The least among you.
- Join Date
- Apr 2005
- Location
- Jacksonville, FL
- Posts
- 981
In that case, your request seems pretty reasonable. It's a shame you had to go to WHT to twist their arm.
Now selling BigVPS's!
Jacksonville Colocation and dedicated servers by colo4jax
We are *not* a reseller. We own our servers, switches, routers and racks.
-
12-11-2006, 07:23 AM #6Web Hosting Master
- Join Date
- Sep 2004
- Posts
- 760
When someone complain with LT here,I would found LT suggest them for server reload at any case,seem LT support team like server reload in the extreme
-
12-11-2006, 07:24 AM #7Web Hosting Guru
- Join Date
- Apr 2006
- Location
- Mandaluyong, Philippines
- Posts
- 316
Originally Posted by otamendi
If the FSCK completed, and the drive still reports errors, you probably want a new drive totally. This could be why they offered to "clone" it, however that's just copying the same corrupt FS onto known good media.
The "Reload" is either :
1 - Install a new drive + OS so the machine can boot or
2 - Wipe the bad drive clean and re-install the OS
Which one did they do?
So far, you don't know if the FSCK was allowed to complete or if they decided it wouldn't be a good idea for it to complete.
Either way this falls out of the realm of what they need to provide .. aside from them not being a bit more specific.
If you can, have them format a new HD with jfs instead of ext3 (far less prone to errors), slave the old (corrupt) drive so that you can attempt a restore yourself.
I may note their level 1 support isn't the swiftest available .. so be specific about your needs and be sure the tech understands your question.
HTH
-Tim
-
12-11-2006, 07:35 AM #8Web Hosting Guru
- Join Date
- Apr 2006
- Location
- Mandaluyong, Philippines
- Posts
- 316
I may also note, LT provisions servers (by default) in a very, very bad way unless you request something different.
Its not that they go out of their way to do this, they simply allow the default OS installer to lump everything into one big ext3 partition on the drive (default for most Linux installers).
This means, the whole drive (rather than just one or two partitions) is prone to failure, especially with ext3.
The OS installers do this because they have to be usable to someone first trying Linux. LT is a data center and should know better. They do, however provide you a place on the order form for special provisioning requests. I'm not just picking on LT, every other data center (well, most of them) do the same stupid thing.
This is a much better provisioning scheme
A few GB (ext2) to /boot (partition 0)
10 GB (jfs) to / (Holds the OS and then some)
15 GB (jfs) to /var (mysql is going to live here)
15 GB (jfs) to /usr (installed applications typically live here)
100 + GB (jfs) to /home (ext3 if using cpanel)
2 - 4 GB (ext2) to /tmp
2 times the RAM in the box for swap
Personally, I like to setup /, /boot and swap as ext2 or jfs, and use LVM to manage the rest of the space. (formatting each volume as I like). This means if /home is nearly full but I have 8 - 10 GB unused space in /var, I can make adjustments (if using ext3).
Note, JFS won't support LVM shrinking.
This means, if the jfs file system on /var fails, /home, /boot, etc remain (usually) untouched.
If the ext3 file system in /home fails, boot (and others) usually remain untouched.
By breaking things down into more partitions, you *really* limit the amount of damage that can occur due to inodes crapping or an unexpected reboot, etc.
Each partition does NOT need a journaling file system such as ext3 or jfs. In fact you'll boost performance using ext2 for partitions that remain mostly static or have no business journaling (such as /tmp).
You may do better off to pay the fee and bring this up with one of their level 3 techs, so you can be sure your server is provisioned for you in the best possible way. Don't let them lump everything into 1 big ext3 partition, or this *will* eventually happen again.
Hope this helps,
TimLast edited by tinkertim; 12-11-2006 at 07:44 AM.
-
12-11-2006, 07:38 AM #9WHT Addict
- Join Date
- Jun 2005
- Posts
- 127
Thaks a lot it is helping.
They did this one:
1 - Install a new drive + OS so the machine can boot
But the server was working ok they recommended the OS reload only because they couldn't clone the bad drive into a new one i didn't know that i was getting ready to resetting my server all together.
Now they are telling me the new drive is in and the old one is also connected problem is i used to have 2 hard drives and they are not telling me which one was the one failing and if now i have the 3 connected.
Am looking for someone right now to restore my info since i've been looking here and it seems am not going to get any help right away from them so i contacted serverwizards but i guess they are sleeping so anyone know any professionals who could do this right now?
thanks.Last edited by otamendi; 12-11-2006 at 07:49 AM.
-
12-11-2006, 07:53 AM #10Too smart for her own good.
- Join Date
- Feb 2004
- Location
- Your Screen
- Posts
- 3,999
Originally Posted by otamendi
If you are not sure which drive is which now, simply run a SMART-check on your server. The SMART utility will tell you which drive has bad sectors. Then you know which drive is bad.
Does LT owe you moving your accounts? No. The scope of their liability in the event of hardware failure is outlined in their TOS/SLA's. They will give you a new drive for free, and they will offer to clone your bad drive (which is stupid in the case of failure), but that's it. Anything beyond that, you are responsible for.
If your sysadmin is truly a sysadmin, restoring your accounts to the new drive should be easy for him to do ... he should also know how to figure out which drive is bad, it's really not rocket science ???
Otherwise, if you are using cPanel, you could sign up with Platinum Server Management and they will be happy to configure your server & restore the accounts. They only work on cPanel servers though.
BaileyLet's Connect on Twitter! @thatsmsgeek2u || Fighting mediocrity one thread at a time.
-
12-11-2006, 07:55 AM #11Web Hosting Guru
- Join Date
- Apr 2006
- Location
- Mandaluyong, Philippines
- Posts
- 316
Originally Posted by otamendi
This will tell you what drives are in the system, and partitions therein.
Also the output of 'mount' would be helpful. This will tell you what is mounted and what the server used to boot.
From that you can easily discern what is going on in your box. I highly recommend these guys for data recovery :
http://www.rfxnetworks.com
I don't know if they have a 24/7 shop .. but worth a try. ext3 partitions are particularly hard to recover, there are tools available that I can recommend but you really need to have some experience using them.
Those guys have plenty of experience doing just that.
It could be that a simple fsck will fix things to a point where you can just mount the old drive and copy (most of) the data over.
/var , /etc , and /usr are crucial for this to happen. (not to mention /home , but users tend to have their files backed up locally and can always upload again).
If not, do you have your backup tarballs ? FSCK could also make those available for you.
If the drive is as bad as it sounds, I would not do *anything* until someone has looked at it for you. FSCK at this point could also hinder other recovery attempts.
Good luck on this one, nobody likes it when this happens
-
12-11-2006, 07:58 AM #12Web Hosting Guru
- Join Date
- Apr 2006
- Location
- Mandaluyong, Philippines
- Posts
- 316
Also is it 2 drives, or 3 ? I've seen you mention both.
-
12-11-2006, 08:03 AM #13WHT Addict
- Join Date
- Jun 2005
- Posts
- 127
Support just told me it has only 2 slots for hard drives my server so, only the new one and the main old one are connected right now, when i finnish copying my files from the old drive i notify them and they connect the old secondary one which is i suppose working ok.
I am now looking for someone to do the copying i have sent emails with details to:
serverwizards
firmbit
rfxnetworks
and rack911
Still nothing from them but i guess they are going to know what to do, am waiting on a first comes first serves basis jejeje with a nice price of course.
Thanks to all.
-
12-11-2006, 02:21 PM #14WHT Addict
- Join Date
- Jun 2005
- Posts
- 127
Well everything seems to be back to normal.
Serverwizards did it for me, they came, they saw, they did it.
Restore all of my accounts back into the new hard drive.
thanks for the help to everybody.
-
12-11-2006, 02:47 PM #15Disabled
- Join Date
- May 2006
- Posts
- 1,426
hardware failure is their fault, so is replacing it like it was when you bought it. However you are responsible for your own backups they shouldnt have charged you a fee for this service