larry007
The first thing I would do is STOP playing with the system and making changes etc etc etc.
Immediately disconnect both drives and secure them safely for later analysis. If there was data to be recovered then due to your actions carried out already, it may be too late to recover.
Ideally you would clone each drive and put the originals aside for later data retrieval.
Using the cloned drives I would try to boot up to ONLY one drive at a time in degraded software RAID mode.
That way you will see what configuration and data is on each drive. Don't make any changes, just observe.
You mentioned a problem booting to the slave, you should be able to do this IF THE DRIVE IS OK, you may need to configure the BIOS to boot from hdc ( D: ) instead of hda ( C: ). I was just doing this the other day and that's what I had to do. The BIOS setting was stopping the Secondary channel drive from booting alone in degraded mode.
Perhaps due to the wrong boot kernel, your data was only being written to one drive, and that drive has failed, and the system reverted to the other drive which had old data on it (as it was hda and the system was only booting to that drive).
Perhaps a user swapped a drive, and replaced it with a drive that had old data on it and the system is now booting up to that drive, is that possible ?
I think drives should be clean when added to an array, or you risk booting to the OS and/or config on the newly inserted drive, rather than the current data.
If both drives are now sync'd to old data, then there will be nothing to retrieve from the second drive in the array. If the array is not sync'd then you stand a chance of getting something.
All your incorrect config issues could be explained by the above suggested scenarios.
It is also possible that a user inadvertantly did a system restore, I don't think sme is capable of automatically doing that. Is it possible for users to do that in the situation the server is in.
> Emails simply came through again as unread - interesting feature....
To be expected if the message files are restored to the Maildir cur folder
> Work files and data, including financials of more than a month is simply > missing which is totally unacceptable, not even for Windows! What can I > do about the data that simply dissapeared?
Clearly your system is reading old data, as that is all that is on the array disk(s). There is no current information on the disk or disks being read.
> The server was rebooted in my absence, presumably with the reset button....
So you don't really know what someone did, and they probably won't tell you either. Therefore anything is possible.
> From the console I re-configured the server and all was now working as it should.
A sign that old config existed on the server.
> But how it is possible .... that data is merely lost.
Lost is probably not how it happened, replaced may be a better word to use.
Hopefully if you can boot to the "other" drive in the array, it may have the more recent data on it.
Unfortunately if any of the above suggestions do not work out, then your ONLY answer is to restore from recent backups. You said something about no USB disks, so are you saying that this company does not do backups at all, for a whole month too ?
If so, I bet they will (do daily backups) after this !
19 Replies
3210 Views