Promise RAID question (sx4000)

  • Thread starter Thread starter Markus Weigand
  • Start date Start date
M

Markus Weigand

I currently have a raid5 with 4 samsung 120GBs running on a promise
Sx4000. Now one of them died and is on its LONG LONG way back to
whereever samsung drives are born (3 weeks and counting), and because
nothing really important is currently in the array it is running in
critical state (until the replacement disc arives).
No problems occured, but today i tried to copy some stuff to the array
and there came a messsage "not enough free space on drive g:", although
chkdisk, explorer and the volume manager report 116GB free space.

Neither a scandisc run nor a defrag run have shown any problems...

Has anyone an idea what the reason for this behaviour could be?
 
Previously Markus Weigand said:
I currently have a raid5 with 4 samsung 120GBs running on a promise
Sx4000. Now one of them died and is on its LONG LONG way back to
whereever samsung drives are born (3 weeks and counting), and because
nothing really important is currently in the array it is running in
critical state (until the replacement disc arives).
No problems occured, but today i tried to copy some stuff to the array
and there came a messsage "not enough free space on drive g:", although
chkdisk, explorer and the volume manager report 116GB free space.
Click to expand...
Neither a scandisc run nor a defrag run have shown any problems...
Click to expand...
Has anyone an idea what the reason for this behaviour could be?
Click to expand...

Lets see. 120GB * 3 = 360GB. 360GB - 116GiB = 235GB = 219GiB.
Hmmm. Does not look special.

RAID-5 in degraded mode looks the same to the OS as RAID-5 normal.
At least it should. In addition, 116GiB = 124GB, which is larger
than the missing drive.

A filesystem-limitation? Maybe a limitation on the number of files?

Arno
 
Arno said:
Lets see. 120GB * 3 = 360GB. 360GB - 116GiB = 235GB = 219GiB.
Hmmm. Does not look special.
Click to expand...
I first thought that maybe the controller rebuild the array from a
critical 4 disc to a non critical 3 disc array, but there are no
indications in the log-files and i think if it were possible it would be
on the feature list.
RAID-5 in degraded mode looks the same to the OS as RAID-5 normal.
At least it should. In addition, 116GiB = 124GB, which is larger
than the missing drive.
Click to expand...

But it is suspiosly near the size of one drive....
A filesystem-limitation? Maybe a limitation on the number of files?
Click to expand...
Just checked. there are about 250k files on the array, which should be
no problem for NTFS.
Also notable is that there are no entries in the system log, neither by
the promise message agent nor by the windows fs driver.
 
Previously Markus Weigand said:
I first thought that maybe the controller rebuild the array from a
critical 4 disc to a non critical 3 disc array, but there are no
indications in the log-files and i think if it were possible it would be
on the feature list.
Click to expand...

It cannot do that. In order to do so it would need to understand
and resize the filesystem on the disks. However to a RAID controller
a HDD is just a sequence of sectors and nothing else. In addition
the whole array should be presentet to the OS like it was a single
disk, regardless of whether it is degraded or not.
But it is suspiosly near the size of one drive....
Click to expand...

Agreed. But that has to be coincidence. This is very likely not something
caused by the missing disk. Or at least not in any obvious way.
Just checked. there are about 250k files on the array, which should be
no problem for NTFS.
Also notable is that there are no entries in the system log, neither by
the promise message agent nor by the windows fs driver.
Click to expand...

Very strange. Sorry, no idea. MS-stuff is not my strong side, since I
seldomly use it. You could test some things, like replacing some
small files with large ones, but that will not give fundamantal
insights, just collect symptoms.

Maybe somebody else here has an idea?

Arno
 
Back
Top