• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 321
  • Last Modified:

Preparations to reinstall SBS Server 2003

The C: drives on this server failed their SMART test and are in a predicted failure state.  I have replacement drives on the way but am trying to think of what I need to do to prepare for this reinstall.

I have both a data and a systems state backup but will do new ones this evening.

I have been backing up the exchange database by using the builtin Backup software and backing up this directory:
SERVER2003\Microsoft Information Store\First Storage Group

On this network we don't use Exchange to handle the mail, only the public calendar and the public contacts.  Am I backing up the correct data for Exchange?

What else should I do to prepare?

Thanks a bunch!

0
wndata
Asked:
wndata
  • 5
  • 4
  • 3
2 Solutions
 
rindiCommented:
A server uses raid, right? Just replace the first drive and wait for the array to get synced. When finished, repeat this with the 2nd drive and so on. You shouldn't need to reinstall anything.

Disk Image backups are usually the fastest and easiest way to make quick backups before meddling with hardware, and they are easiest to restore if things go bad. Use bootit-ng for that, the trial version works without restrictions.

http://terabyteunlimited.com/
0
 
Alan Huseyin KayahanCommented:
Hello wndata,
    As rindi mentioned, if your server is configured for RAID 1, follow directions from rindi.
     If not, backing up sys state and exchange information store would be sufficent to have Public folders back.
   I have two suggestions.
          One is to take full backup of the entire server and restore it to new disk. That would be better for any installed programs etc, or sharepoint settings and so on. SBS is IIS sensitive.
         Other one is, again take a full backup to use just in case (or use BootItNG as rindi suggested to take backup), then insert the new HDD without removing the current one, then in RAID configuration utility, create a RAID 1 so new disk will be an exact mirror. If you are unsure how to do this without erasing the data in main disk, then use windows disk management, convert both disks to dynamic, create the mirror. If current HDD has enough life to complete the process and mirror is created, remove the bad disk. No need to restore, partition managers, install from scratch etc. Less time consuming.

Regards
   
   
0
 
wndataAuthor Commented:
Thanks folks for responding.

This server is RAID 5.  There are two mirrored drives functioning as C: and three data drives - data drives are healthy.  Initially I thought  I could do as Rindi suggested and replace one of the C: drives and wait for it to complete the mirror and then replace the other but I have been talking to Dell techsupport and they are saying I should reinstall.

The tech said,  since both mirrored drives are showing the same error, that the error was propagated from one drive to the other and I should reinstall.

Obviously I would MUCH RATHER take advantage of the mirroring and avoid the rebuild.  Do I have good information from Dell and have to rebuild?

Would it hurt to try the mirror and then, if necessary reformat the new drives and reinstall?  

Another consideration:  The C: drives on this server are partitioned for a root directory of 12 GB (I didn't set it up originally).  I have been doing a lot of housekeeping to keep C: with enough space to function well.  I understand that software partition managers for servers don't work very well.    So - if I try the mirroring and it works, I am still stuck with 12 GB C: drive.  Is there a way extend the partition safely?

 I haven't had to deal with this situation before and I certainly would appreciate experienced input.
0
Transaction-level recovery for Oracle database

Veeam Explore for Oracle delivers low RTOs and RPOs with agentless transaction log backup and transaction-level recovery of Oracle databases. You can restore the database to a precise point in time, even to a specific transaction.

 
rindiCommented:
A SMART error isn't something that gets propageted. A Smart error means your drives are still working, but they are expected to break down soon. so as long as react promptly there should not be any issues with your data. That is the point of SMART, it is a system that warns you beforehand of coming errors. So changing the HD's one by one should not be a problem at all.
0
 
Alan Huseyin KayahanCommented:
   I would first suggest you to install a SMART monitor software, usually you can download that utility from the web site of your hdd manufacturer such as seagate etc, or one from following link. I use Active SMART but its not freeware.
http://www.freedownloadscenter.com/Best/hdd-free-smart.html
   Most of the SMART issues I face generally caused by high temperature. As mentioned, SMART issues does not get propagated, it looks like that way, but in fact, it is a common thing that affects all HDDs such as temperature. But since your R5 array is not affected, it also may be a controller issue, generally followed by "delayed write failed" errors in windows.
    Yes you will be stuck with same partition sizes. But BootITNG is an excellent partitioning program. You can resize your active partition easily. I would suggest resizing the new healthy disk after mirroring completes, resizing (moving empty space etc) then mirroring sector by sector may kill the current disk if SMART issue is serious.
0
 
wndataAuthor Commented:
Thanks again for the input.   This server is in an upstairs room and it gets warm up there - not beastly but uncomfortable in the summer.  I would guess that it gets at least 80 for 5 months or so of the year.  I have heard the fans rev up and wondered about the heat.  Is 80 degrees enough to diminish the hard drive life?  I don't know why they put it up there in the first place but, as I said, I wasn't in the loop when the server was installed.  

MrHusy -  You didn't say but I am guessing that the hdd-free-smart software will give me more detailed information about the origin of the SMART errors?

   
0
 
Alan Huseyin KayahanCommented:
" am guessing that the hdd-free-smart software will give me more detailed information about the origin of the SMART errors?"
   Exactly. But for real life, installing SMART software and spotting the issue "after having smart warnings" have never helped me to return the disk to a stable state. Most of the times it was temperature. You do extra cooling and SMART error disappears. But lets say that your servers run in inappropriate temp for long time (Temp is high in room<Higher in server case<highest in HDD). That high temp already affected your disk health. So making error disappear has no recovering effect on disk health. Most probably delayed read write errors, event id 32 (if i remember correct), instant hanging mouse freezed, are on the way. SMART software is to install and monitor disk health with tresholds before you get SMART warnings.
      I dont want to digress, your initial question's answer is given correctly by rindi
"Just replace the first drive and wait for the array to get synced. When finished, repeat this with the 2nd drive and so on. You shouldn't need to reinstall anything."


0
 
wndataAuthor Commented:
MrHusy

I followed the link  you provided but there are several choices that look, to me, like they might be appropriate.

Which one were you recommending:

DiskCheckup 2.0  - this one mentiones S.M.A.R.T.

SMARTHDD 0.8.0.7440  

HDD Observer 3.1  

Thanks again!
0
 
rindiCommented:
RAID controllers mask the HD's to the OS, so installing a SMART utility will probably not help you get more info on the disks, as such tools require direct interaction with the controller and HD's.

If you want to use such tools you would have to connect the HD to a non-raid controller and then use the utility.
0
 
wndataAuthor Commented:

Oh okay.  One last thing - I don't imagine that swapping out the raid 1 drives is as easy as turning off the server, removing one of the affected drives and replacing it with a new drive so could you point me at some step by step instructions?

I really appreciate your help!
0
 
rindiCommented:
Most servers have hot-swapping raid controllers. You should check the manual of your server, and if it has hot-swap, you just remove the bad disk while the server is running, then insert the new HD and the rebuild should automatically initiate. This should also be mentioned somewhere in the users manual.
0
 
wndataAuthor Commented:
Thanks to Rindi AND MrHusy.  They both assisted me.  Just thought you would like to know that the rebuild did work even with the predicted fail codes on both drives.  I was REALLY pleased to save all the time and effort that would have been spent on the reinstall!  Thanks to you both!  I will be downloading BootitNG and resizing the C: drive next.  

0

Featured Post

What does it mean to be "Always On"?

Is your cloud always on? With an Always On cloud you won't have to worry about downtime for maintenance or software application code updates, ensuring that your bottom line isn't affected.

  • 5
  • 4
  • 3
Tackle projects and never again get stuck behind a technical roadblock.
Join Now