Thorz
11-04-2007, 08:54 AM
Hello
This is my first RAID 1 build and I have done my homework and read lots of info before posting.
I wanted to write enough organized background information on this post together with my questions. As it came a little long I have marked my 3 questions in red. I like to thank you all in advance for trying to help me.
I have built my last machine using Windows XP Pro with the following storage setup:
On the ICH9R controller:
SATA1: Raptor 74GB -> OS
SATA3: Seagate 1TB -> No RAID
SATA4: Seagate 1TB -> RAID 1 (disk A)
SATA5: Seagate 1TB -> RAID 1 (disk B)
SATA6: Seagate 1TB -> No RAID
On the jMicron eSATA ports:
eSATA1: WD 1TB -> No RAID
When I setup Windows I used IDE mode on the ICH9R. Everything was working smoothly. OS recognized and full formatted all the disk. I got of course BSOD after changing to RAID mode on the BIOS, but luckily this board has those jMicron eSATA ports. I connected the Raptor with the OS on one of these and could install the Matrix Storage Manager downloaded from Intel. After putting the Raptor back on the ICH9R port XP booted without any problems. I then created the RAID 1 from scratch from inside Windows and all was good. Here is where the things started to go bad:
1. I got an error every time I tried to full format the RAID 1 from XP's Disk Manager. After some minutes into the formatting process I just got a warning from the Matrix Storage Console that BOTH my RAID 1 drives have failed! I got the same message when booting. I tried rebuilding the RAID set and recreating it but nothing helped, I always got the fail in the middle of the full format. I then decided to use a quick format instead and things appeared to be ok. I loaded the RAID set with 650GB of info for testing and it appears to be working fine.
I was left very uneasy about this fail. I am going to store important info on this RAID and a future fail could be very annoying.
QUESTION:
Do you know the reason for that the RAID 1 set cannot be full formatted? Should I be worried? Any comments/ideas are welcome.
2. After checking the options on the Intel Matrix Storage Manager I found that "Hard Drive's Write Cache" is enabled for the "array_000" but found an option to enable the "Volume's Write-Back Cache". This was confusing. The help states that the last one is just for RAID 5 sets.
I went and check the properties of the RAID volume under Disk Manager and found under Policies that "Enable write caching on the disk" was not enabled. The 2 options for "Write caching and safe removal" were grayed out but I can see that "Optimize for performance" is selected. I tried to enable the write cache but I got this entry on the Event Viewer:
EVENT ID: 34
The driver disabled the write cache on device \Device\Harddisk1\DR1.
After this, the option "Enable write caching on the disk" that I had just marked on Disk Manager became grayed out. The same thing happened for every disk connected to the ICH9R after I tried to enable the Write Cache from Disk manager on them.
BENCHMARKS RAID 1:
SANDRA physical: Drive index of 73MB/s
HD Tach Average Read: 87.9 MB/s.
These values are almost identical to the results of the Non RAID (same type Seagate 1TB) disk connected on the SATA3 port of the ICH9R. The SANDRA scores are also identical to the same disk on SATA3 when moved to a jMicron eSATA port.
The only results that were different were:
HD Tach's "Burst Speed":
Non RAID: 231.9 MB/s
RAID 1: 139.5 MB/s
ATTO v2.34 (0.5-1024KB Total Length 32MB, Max Values):
Non RAID: Write: 83506 Read: 87257
RAID 1: Write: 95189 Read: 118987
QUESTION:
Is this the normal behavior of the disks connected to the ICH9R? I don't understand why the Burst Speed of the RAID 1 volume is so low compared with a single disk connected to the same controller.
3. 2 of my Seagate disks have no fan installed in front of them. One of them is on 53C. The ones with fan installed (on the bottom of the case) are reporting ~35C.
QUESTION:
Is 53C a dangerous temperature for these disks?
I appreciate your answers and opinions. Thanks a lot.
This is my first RAID 1 build and I have done my homework and read lots of info before posting.
I wanted to write enough organized background information on this post together with my questions. As it came a little long I have marked my 3 questions in red. I like to thank you all in advance for trying to help me.
I have built my last machine using Windows XP Pro with the following storage setup:
On the ICH9R controller:
SATA1: Raptor 74GB -> OS
SATA3: Seagate 1TB -> No RAID
SATA4: Seagate 1TB -> RAID 1 (disk A)
SATA5: Seagate 1TB -> RAID 1 (disk B)
SATA6: Seagate 1TB -> No RAID
On the jMicron eSATA ports:
eSATA1: WD 1TB -> No RAID
When I setup Windows I used IDE mode on the ICH9R. Everything was working smoothly. OS recognized and full formatted all the disk. I got of course BSOD after changing to RAID mode on the BIOS, but luckily this board has those jMicron eSATA ports. I connected the Raptor with the OS on one of these and could install the Matrix Storage Manager downloaded from Intel. After putting the Raptor back on the ICH9R port XP booted without any problems. I then created the RAID 1 from scratch from inside Windows and all was good. Here is where the things started to go bad:
1. I got an error every time I tried to full format the RAID 1 from XP's Disk Manager. After some minutes into the formatting process I just got a warning from the Matrix Storage Console that BOTH my RAID 1 drives have failed! I got the same message when booting. I tried rebuilding the RAID set and recreating it but nothing helped, I always got the fail in the middle of the full format. I then decided to use a quick format instead and things appeared to be ok. I loaded the RAID set with 650GB of info for testing and it appears to be working fine.
I was left very uneasy about this fail. I am going to store important info on this RAID and a future fail could be very annoying.
QUESTION:
Do you know the reason for that the RAID 1 set cannot be full formatted? Should I be worried? Any comments/ideas are welcome.
2. After checking the options on the Intel Matrix Storage Manager I found that "Hard Drive's Write Cache" is enabled for the "array_000" but found an option to enable the "Volume's Write-Back Cache". This was confusing. The help states that the last one is just for RAID 5 sets.
I went and check the properties of the RAID volume under Disk Manager and found under Policies that "Enable write caching on the disk" was not enabled. The 2 options for "Write caching and safe removal" were grayed out but I can see that "Optimize for performance" is selected. I tried to enable the write cache but I got this entry on the Event Viewer:
EVENT ID: 34
The driver disabled the write cache on device \Device\Harddisk1\DR1.
After this, the option "Enable write caching on the disk" that I had just marked on Disk Manager became grayed out. The same thing happened for every disk connected to the ICH9R after I tried to enable the Write Cache from Disk manager on them.
BENCHMARKS RAID 1:
SANDRA physical: Drive index of 73MB/s
HD Tach Average Read: 87.9 MB/s.
These values are almost identical to the results of the Non RAID (same type Seagate 1TB) disk connected on the SATA3 port of the ICH9R. The SANDRA scores are also identical to the same disk on SATA3 when moved to a jMicron eSATA port.
The only results that were different were:
HD Tach's "Burst Speed":
Non RAID: 231.9 MB/s
RAID 1: 139.5 MB/s
ATTO v2.34 (0.5-1024KB Total Length 32MB, Max Values):
Non RAID: Write: 83506 Read: 87257
RAID 1: Write: 95189 Read: 118987
QUESTION:
Is this the normal behavior of the disks connected to the ICH9R? I don't understand why the Burst Speed of the RAID 1 volume is so low compared with a single disk connected to the same controller.
3. 2 of my Seagate disks have no fan installed in front of them. One of them is on 53C. The ones with fan installed (on the bottom of the case) are reporting ~35C.
QUESTION:
Is 53C a dangerous temperature for these disks?
I appreciate your answers and opinions. Thanks a lot.