Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations SkipVought on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

increase size of raid5 4

Status
Not open for further replies.

mennonitevillage

IS-IT--Management
Jun 10, 2003
16
0
0
US
fairly new to raid need to know:

1: if I swapp one of the drives in raid with a larger drive will the raid replicate the data?

2: how do I setup an online spare?
any help would be appreciated
 
1. Yes, but never do this when the server is on and the drive state is NOT failed. If you replace one non-failed drive it should be done with the power off. Doing this will NOT give you extra storage space until you replace all of the drives with ones of larger capacity.

Turn off the server, replace one drive with one of larger capacity, boot the server and wait for the Array to rebuild. Repeat until all drives are replaced with larger drives, or just make a good backup, power down the server, erase the array, replace all of the drives with larger ones, restore from backup.

2. You can create a online spare when you create an array or when you add an extra drive that's not part of an array.

Always read the manual for your specific SMART Array Controller, as different models have different capabilities.
 
The server does not have to be down to swap a live disk. There is a risk swapping a live disk out, but that is what RAID is designed for so the server should handle it well.

I have swapped out many disks for larger ones, but as stated all disks have to be swapped before you can use the extra diskspace. Swap one at a time, ensure the restriping has completed, then swap another one.

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------
 
You're taking a huge risk pulling out Non-Failed disks while they're actively reading/writing. That's not what they're intended to do. Hot-plug is intended for replacing failed drives, that the controller is not actively writing to.

Power down the machine first, safe is better.
 
No bigger risk than a drive failing as it's busy writing - the controller should handle it.

That said if you're really scared, do it at a low activity time (if there is one) far safer than repeatedly power cycling your server.
 
just remember that once the larger drives have been put in you will need to create a new logical drive using the array config tool as expanding the existing logical drive will trash the data.
 
thanksthe point is trying to increase the size of the raid,
initially set up 60 gig for user/data drive for whole domain eith saving databases, photos, ect...

were starting to runn out of room.

what are some other alternitives?
 
The alternative is to delete the array, recreate with the new drives then restore your data. You could also create a new array and mount it in the original volume using Windows.

Enforcing some kind of quote is also useful. I like to scan my volumes for large files, i.e. anything over 25MB to see what I find. Sometimes you find gigantic temp files that you can delete, or someone's MP3/WMA/Picture collection that's using up your space.
 
Building on patrickrouse's ideas - I actually scan regularly for files called
*.VIR (I don't want quarantined virus files on my server)
*.TMP (I don't want a pile of temporary files either, if I can't delete them it's because they're in use and I don't mind that so much)

If you have users using Lotus Notes with files on the server, take a look for CACHE.DSK - this file can be huuuuuge sometimes for no reason and if you delete it, it should be recreated next time Notes runs...... your mileage may vary on this one though [pipe]

I know that Netware starts to suffer performance issues above around 80% volume utilisation, and it starts to get really critical above 90-95% so you can prevent your performance from being impacted by limiting the total space in use to <80% of the physically installed drive capacity. As it's not a precipitos fall off in performance you can release some of the space in an emergency and the users will probably not notice.

By applying that limit - it also gives you breathing space to be able to plan an upgrade because you have 20% held in reserve that the users don't know about that you can release to counter an immediate problem, which means you will rarely have to make a panic upgrade or purge on your [love] anniversary when you've promised the missus that you won't be home late but the drive has filled up and payroll won't run unless you fix it etc. [thumbsup2]
 
NetWare volumes also have the capability to selectively compress files based on a compression factor, i.e. only compress when the file can be compressed 75% otherwise leave alone.

Since migrating to Windows Servers from NW4.11 I never use compression, as I like to know exactly how much storage space I'm actually using, how much I need to backup, and I don't have to worry about performance of large files uncompressing that may fill the disk.

Itry to keep at least 40% free space on every volume, which lessens the chance of accidently filling a disk with log files, temp files.... on top of that I've never seen the need for disk space decrease, so when it goes below 40% I either archive old files to DVD (i.e. anything that hasn't been accessed or modified in 36 months) or ask for money for more disks, whichever management is more comfortable with.




 
to TheLad,
If I shut down then swap out 1 drive boot up let the raid finish the restriping then shut down repeat till done is this safe?
Will I be able to use the larger disk space with out having to:
&quot;just remember that once the larger drives have been put in you will need to create a new logical drive using the array config tool as expanding the existing logical drive will trash the data&quot; It doesn't seem logical, if you have to replace the drive what if you can't find the same drive anymore, if you were going to need to recreate the raid it seems to make more sense to have a two large drives mirrored

just a thought
 
Admin Tip...

Great Admin utility....and it's free!
Most of this thread discusses space consumed by the users, so with that thought...figured I'd share this freebe:


It's the cleanest, no-install drive usage display program I have seen.

But on the current topic a full restore to a clean raid is the only way to go....you don't need to ask for problems and if you loose the raid you have one big problem!

Good Luck....Ron
 
I have a need to do something similar.

I have a raid5 array made up of 18GB drives. I want to replace those with 72GB drives? This array is a backup drive so I don't think anyone will freak out if the data is lost.

Is the best way - power off server, replace one at a time, then I am done?

OR

Erase the array , put in the drives, create a new array?
(and if its the second one, how do I do it exactly??)
 
I wound up backing up all the files to a temporary large drive on the network and creating a clean new raid in the raid controller then after booting up with the new raid formated the drive(do not make it a dynamic partition)then restored the filesafter the raid was built.
 
The first option (individual disk swap) is good for no/hardly any downtime. Erasing the existing array would cause a total loss of your system - that is if your operating system is stored on the existing RAID5. If the RAID5 is just the data store and the OS is stored on other disks, you can perform the procedure without a total loss.

If you have the time, go for the complete new server rebuild. If not, swap the disks individually, wait for the RAID to restripe, then swap another disk. Recommendation: try do this at a time when the server is less busy. The extra diskspace can then be created as a new Logical Drive and allocated accordingly.

-----------------------------------------------------
&quot;It's true, its damn true!&quot;
-----------------------------------------------------
 
OK - do you:

delete the array first, power down, replace the drives, power up, create the array, create the partition, done?

Or - do I have to delete the partition in Windows first?

 
After copying the files of the shared raid, I rebooted then hit f8 when the Compaq smart array controller came up.

I then selected delete logical drive, select the raid 5 (not the os), after deleting the raid I saved shut down replaced the drives. I powered up, f8 again, create logical drive, this time create raid 5 rebooted and then started up windows.

After I logged into windows I went to Computer Management Console selected disk management think it started the dynamic disk new Volume wizard. I canceled, right clicked on the gray box with disk 1 and selected convert to basic disk, after that I right clicked on the unallocated space and hit create new partition.

This may not be the best way but was simple and easy
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top