Topic: Recovering from corrupted Hardddisk to new one

[mazkot]

hi,

I would like to ask if this is plausible.

Background:

I have a sme installed in a 160GB harddisk. It is running as webmail, fileserver, and web server. with 20 users for webmail and 6 heavy users of fileservers
last week our request to upgrade the HD to 500 gb was approved. Sat new 2 new 500gb hd arrived. emailed all users to back up all their files to their local computer and specifically drive d so I could backup and  upgrade tonight.

Monday morning for unknown reason server stopped. tried rebooting but I only get to "enter runlevel prompt"

run sme rescue and got the error that disk full. so checking the disk found 100% used.

the question:

since I cannot run the console and do the back up can I:
- install new Sme in new 500gb with raid
- attached old hd and follow the instructions of:
   http://wiki.contribs.org/Backup_server_config
So all files and users are exported.

Is is possible to do this?

thanks

[janet]

mazkot

Run sme in rescue mode from CD and mount the disk.
Then remove some unwanted files or move them off to USB or wherever.
Then start your server normally.
Follow this, remembering to do the pre backup events to dump the databases.
http://wiki.contribs.org/UpgradeDisk

[Stefano]

run sme rescue and got the error that disk full. so checking the disk found 100% used.

my bet: your backup failed to mount the external usb disk/remote share and did the backup on your HD..

[mazkot]

thanks for the reply. Still waiting for their decisions on which files to be deleted. But if they could not decide is may plan doable or I'm just going to get burn in doing it?

for now might start to look at files that was added from sat to monday..

[janet]

mazkot

...is may plan doable or I'm just going to get burn in doing it?
......attached old hd and follow the instructions of:
   http://wiki.contribs.org/Backup_server_config

Well it depends which part of that article you are going to follow, you need to restore all required files, not just users & ibay data.

[mazkot]

hi mary,

i was planning in just doing the:
1. Attach the 160 gb directly to the restored machine with new sme
and do the following from the wiki.
1. Copying data to the restored server
2.Manually transferring configuration information


but as I re read it again it seems it need two to tango. Since I cannot issue the commands in the old server : "signal-event pre-backup"

or I can just go ahead and try it.

[janet]

mazkot

You asked if the process you proposed doing would work or not. I responded:
"Run sme in rescue mode from CD and mount the disk.
Then remove some unwanted files or move them off to USB or wherever.
Then start your server normally.
Follow this, remembering to do the pre backup events to dump the databases.
http://wiki.contribs.org/UpgradeDisk"

So in order to do the signal-event pre-backup on the old server, you need to mount the drive and remove data, so you can eventually start up that server normally.

If you don't do the signal-event pre-backup then you won't have any mysql database details transferred in the backup/restore process and your restored server is likely to be in an even bigger mess.

You can go ahead and try anything you want, but if you overlook to do correct procedures then don't blame me/us for it not working.

[mazkot]

mary,

I was not trying to be a knuckle head and I have no plans of blaming the group. I am now blaming the management for taking 2 years just to approved the HD upgrade.

I had done what you said I deleted some files giving me an 83% used HD. I cannot move the files to the usb because it was not detected. but the same problem persist after loading I am faced with the "Enter runlevel" command I have entered enter level from 1-6 and have the same outcome

INIT: Entering runlevel:
INIT: no more processes left in this runlevel

As you stated no old server no import of users and files.

So I am left with manually changing the permissions of all users file and mail?

[MSmith]

How about running the drive maker's utility CD to check the hard drive for physical errors, then if it comes up clean, doing an "in-place upgrade" with the SME CD to see if that'll fix the problem and make the server bootable again?  If you really want to be safe, after the drive check you can copy everything from the 160-gig drive to one of the 500-gig drives after booting from a Linux LiveCD, just so you have a copy of the data.

As a bonus, you can try the in-place upgrade by mounting the old drive in any old hardware, thus having a machine from which you can do an Affa backup/rise to the "new" server with the mirrored 500-gig drives.  Easy as can be, and preserves *everything*.

[mazkot]

Thanks all. was able to transfer the it to the new HD.

did the "in-place upgrade" and was able to boot properly

and http://wiki.contribs.org/UpgradeDisk