1

Topic: How long should it take to add 100 GB to a SLES VM

I have a SLES 10 SP3 VMWare Guest on ESXi that had filled it's 350 GB hard drive.  I have used GParted many times on VMs with similar configurations, but smaller disks.  It has always gone without issue, so I had no qualms about using it to add 100 GB to this drive.  In fact, I had another VM on a different ESXi server that I expanded from 20 GB to 50 at the same time.  That took less than 5 minutes.

I started this expansion process at 6:00 AM this morning.  It is now 11:20 PM, and it is still running.  Since I have never stopped one of these processes, I am afraid to do so now.  The data on this drive is very important, and I don't want to increase the risk of loss or damage.

Could it still actually be working?  The slider continues to slide back and forth.  The details have been displaying the reiserfsck command all day.  I plan to let it run through the night in the hopes that it will actually complete successfully.  But, I'm getting nervous that is is really not doing anything.  Is there any way to tell if it is actually running?  Would you expect that it will eventually complete?  I would have expected it to post an error if there was a problem.

I look forward to a response.

2

Re: How long should it take to add 100 GB to a SLES VM

You might try checking process status to see if the reiserfsck is still running.

E.g., ps -ef | grep -i reiser

The reiserfsck command is responsible for checking and fixing a file system.

What version of GParted or GParted Live are you using?

With regards to time, if you added the extra space on to the end of the partition, then it should only take minutes to complete.  If you moved the start of the partition then that can take a long time.  In your case it appears that a file system check is being performed so perhaps there was some problem with the consistency of the file system data.

3

Re: How long should it take to add 100 GB to a SLES VM

Thank you for replying.  It is still running.  33 hrs and counting.
All I did was add space to the end to go from 350 GB to 450 GB.  I've done this numerous times in the past (though not with so large a volume) and never had an issue.

This server, had filled the disk (hence adding space).  It is an eMail Archiving server, so it has a database and lots of little files on it.

When I look at he vSphere Client Performance Monitor, or at Veeam Monitor, it shows activity in that VM, just reading from the disk at a rate averaging 700 KB/s.  The GParted screen still just shows the slider going back and forth, and it hasn't moved on from the Reiserfsck.

I'll be patient and let it keep going.  I was just hoping to get an idea of how long to expect, or a way to verify that I should expect it to complete at some time.  I have not seen any errors posted anywhere.  I am afraid to cancel for fear of damaging something.

  I'd be very interested on input as to what you think canceling at this stage could do, and whetehr you think it sounds like it will eventually complete.

Thank you.

4

Re: How long should it take to add 100 GB to a SLES VM

More activity.  Last night some time, the GParted interface advanced (sort of).  It now states for a second time that it is expanding and checking the disk.  I'm letting it continue because I am not sure what it is really doing or how else to address it without risking damaging the data or losing access to it.

  Here's how it looks now:

The top section still says:
Grow /dev/sda2 from 347.99 GiB to 447.99 GiB

Then the horizontal slider is going back and forth continually.

check file system on /dev/sda2 for errors  and (if possible) fix them

Completed Operations

0 of 1 operations completed

The Details section appears as follows:
  \/  grow partition from 347.99 GiB to 447.99 GiB
             old start: 4209030
             old end:   733993784
             old size:   729784755 (347.99 GiB)

             new start: 4209030
             new end:  943718399
             new size:  939509370 (447.99 GiB)

  \/  check file system on /dev/sda2 for errors  and (if possible) fix them

       reiserfsck --yes --fix-fixable --quiet /dev/sda2

  \/  grow partition from 347.99 GiB to 447.99 GiB
             old start: 4209030
             old end:   733993784
             old size:   729784755 (347.99 GiB)

             new start: 4209030
             new end:  943718399
             new size:  939509370 (447.99 GiB)

  \/  check file system on /dev/sda2 for errors  and (if possible) fix them

       reiserfsck --yes --fix-fixable --quiet /dev/sda2

5

Re: How long should it take to add 100 GB to a SLES VM

By the way, I am using the gparted-live-0.8.1-3.iso

6

Re: How long should it take to add 100 GB to a SLES VM

Just to keep you updated, the process is still running and in the same place it has been during it's second iteration of the process.  I have no idea why it seems to be doing the same thing a second time, but it did it on its own and I am letting it run.

Earlier it was suggested that I check the running processes, but that is impossible.  The process is running in the VM and I booted with the GParted Live CD iso.  So, I can't attach to the server with anything but the vSphere console, and when I click on Terminal in the GParted interface (while the operation is still running) it does nothing.

All I can do is watch the slider bounce and monitor with Veeam and the vSphere Performance tools.  Both of which show Disk Read activity staying consistently between 550 and 700 KB/s.  the last write activity I see was almost imperceptible at 10:09 am on 2/3/2012.

Does this information give any clue as to what is going on?  Does it make sense to keep waiting?  What is the best way to get this server functional again without risking the data?

Thanks in advance for your reply.

7

Re: How long should it take to add 100 GB to a SLES VM

It is still going, and all I can see that it has done in days, is disk read activity.  Some input and direction would be greatly appreciated.

8 (edited by EliotLanes 2012-02-07 15:07:22)

Re: How long should it take to add 100 GB to a SLES VM

Well, what do you know.  98 hrs, but it completed successfully.  I have no idea why it took so long or what it was doing, but I am very glad for the outcome.

I hope this helps anyone else who has a similar experience.  It's nerve racking, but just let it run.

9

Re: How long should it take to add 100 GB to a SLES VM

Thank you for your report.

98 hrs, but it completed successfully.  I have no idea why it took so long or what it was doing, but I am very glad for the outcome.

98 hours is surely one of the longest times I remember in the GParted forum, for a successful operation on a filesystem of this size.

Why so long...  It is possible that the bigger part of this time belongs to a check and fix operation.
This is just a thought. I read in the last few days the following report on the Linux filesystem check:
http://www.enterprisestorageforum.com/s … re-in.html
This comes from a mailing newsletter related to the professional storage systems.
The tests were performed on rather well working (not damaged) xfs and ext4 filesystems (up to 72TB big).
The check times were quite short, up to 1 hour, but some one commented that this check isn't valid for really damaged filesystems, inconsisted data or with hard drive problems. In these cases, the time would be many hours or even days.

I don't know if the reiserfsck version is somehow related to the check time.

*** It is highly recommended to backup any important files before doing resize/move operations. ***

10

Re: How long should it take to add 100 GB to a SLES VM

I think the reiserfsck was what was causing the slowdown, but as far as I can see, there was no problem with the disk or file system.  The only "error" I've ever seen in that regard on this server was the fact that it filled up 100% of the disk space- which is what prompted the re-size in the first place.

I wonder if there is a way you can make the file system check optional- at least the second run of it?

11

Re: How long should it take to add 100 GB to a SLES VM

the fact that it filled up 100% of the disk space

This could be important.
Another factor for the filesystem check & repair is the RAM. Large RAM amounts help.

I wonder if there is a way you can make the file system check optional- at least the second run of it?

I guess the first run was a simulation, the "real" run was the second one.

Parted can't run safely on damaged filesystems, because this can lead to data loss. That's why GParted follows the safest way to proceed.
To ovecome such tests, it is possible to run the specific tools manually from the command line.

*** It is highly recommended to backup any important files before doing resize/move operations. ***

12

Re: How long should it take to add 100 GB to a SLES VM

The comments from class413 are accurate.  We try to ensure data integrity to the best of our ability.  Hence we run file system checks before working on the file system.

We are working on improving the feedback from running commands, but that will not shorten the amount of time the commands take to run.

Ideally we would display an estimate of the time the operation will take.  Unfortunately though we do not know an accurate way of doing this.  Most times operations run quickly, but occasionally like in your situation, a command can take a very, very long time to complete.

If you had killed the command mid-operation, there is a high likelihood of file system corruption.

When earlier I asked for a process status listing, I thought that you might start up a terminal window from your GParted Live session to capture this information.  I should have been more specific at that time.

I am glad that the operation did finally complete successfully.