If the user has created the array without initializing it and reboots the system, the state of the array will be reported as
"New arrays needs init, needs system reboot"
A reboot at this point is not required. The array can be initialized and the array will become fault tolerant.

Pause I/O is a feature which enables replacement of SCSI devices in an operational enclosure that is not specifically designed with hot-swap capability. Although it temporarily stops I/O transactions on the SCSI bus, it cannot assure the same level of protection from electrical noise during device replacement as hot-swap hardware. As indicated in the following procedures which use Pause I/O, only a single change ( such as a device removal OR a device insertion, but not both) should be made with each use of this feature. Do not replace a non-downed drive in a paused array. The proper procedure is to down the drive, and then replace it. Failure to do so may lead to data corruption.

The following are the limitations to the current release:

• For a multiple channel environment, users cannot presently change existing array configurations on the fly through use of Pause/Resume or rescan.
• Geometry of a drive is not updated when a smaller single member is replaced with a larger single member and vice versa. Should this single to single replacement occur, the potential for data corruption is possible.
• Singles cannot be replaced with array members and be reported back to the O/S and application layer properly. This means that in the event of a channel failure a user cannot replace a single with an array member to make his array configuration complete and usable again during runtime.
• Configuration changes of devices from CD to TAPE, TAPE to CD, SINGLE to CD/TAPE, or Array member to CD/TAPE, is not supported.

The software automatically responds to the event of one failed array member by using a spare, if available, and notifying the user. To replace an array member in active use, the following steps must be performed:

1. Down the Array Member Device in all cases.
2. If a hot-swap enclosure is not used, Pause I/O.
3. Remove the Array Member and allow I/O to resume.
   • If a hot-swap enclosure is not used, Pause I/O.
   • Insert the replacement device and allow I/O to resume.
   • It may be necessary to low-level format the removed, downed spare prior to re-use.

To replace a spare drive that is recognized as good by the software, the following steps must be used:

1. Delete the spare.
2. If a hot-swap enclosure is not used, Pause I/O.
3. Remove the Spare and allow I/O to resume.
4. If a hot-swap enclosure is not used, Pause I/O.
5. Insert the replacement device and allow I/O to resume via CI/O AMS.
6. Mark this new device as a spare(dedicated or pool).

To replace a spare drive that is marked as down, the following steps must be used:

1. If a hot-swap enclosure is not used, Pause I/O
2. Remove the Downed Spare and allow I/O to resume
3. If a hot-swap enclosure is not used, Pause I/O
4. Insert the replacement device and allow I/O to resume via CI/O AMS.
5. Mark the new device as a spare (dedicated or pool).

1. When hot-swapping a drive that supports tagged command queuing (TCQ) with a drive that does not support this feature (at the same SCSI ID), a re-scan may not detect the swap of the drive. The system could possibly send commands with tags to the new drive, causing a system hang. The workaround is to perform the hot-swap while the server is down and/or off. Contact the drive manufacturer to determine if the drive supports TCQ.
2. When an array member fails, replace it with a good drive with the same SCSI ID to reconstruct the array. The information about the new array member will not be written to the member's reserved sectors until a manual Reconstruction command is issued after re-scanning the system. (The member might show up as a "single" disk if the system immediately rebooted after the re-scan.)
3. It is recommended that old/down/defective disk drives be removed from the system as soon as possible. Any removed drives that are reused must be low-level formatted before reuse.

The Verify and Reconstruct operations can take a long time to complete. For large arrays, from minutes to several hours. By default, these operations are assigned a Low or Medium priority to minimize the impact to other server activity. For faster operation, change the priority to High.

The use of large fonts is not supported on Windows NT versions of CI/O AMS.

CI/O Array Management Software does not support blinking the drive light on CD-ROM drives.

If all SCSI drives connected to the computer's SCSI channels are members of a single array and that array goes off-line, CI/O Array Management Software does not display any controller, SCSI channels, or drives. If all members of an array are missing, the array and its members are not displayed in CI/O Array Management Software after rebooting or when performing a re-scan.

Citrix Winframe and the AAA-130 series adapter are incompatible.

The installation of Small Business server does not allow to specify additional devices not allowing to add the array1000 family manager's cda1000 driver. To resolve this copy the "winnt.sif file located in the i386 directory of the SBS cdrom to disk # 2 of the install disks. This should allow to specify for additional \ 3rd party drivers.

During NT setup, if the message "Setup is unable to locate the hard drive partition prepared by the MS-DOS portion of setup" or the message "xxxx MB disk y at Id z on bus 0 on cda1000.sys does not contain a partition suitable for starting Window NT" appears, do the following:

1. Re-boot using the ArrayConfig disk.
2. Using ArrayConfig utility, ensure that the array to be used as the NT boot device includes the drive with the lowest SCSI target id on the Array1000 Family products bus.

During NT Setup, the system may hang while the NT Setup floppy #3 is being used for copying the SCSI disk device driver. A workaround is to boot to a DOS boot-disk, create a DOS partition on the array using FDISK and then install Windows NT on the array.

When attempting to boot from NT installation CD, some systems will issue an error message:

		"Boot: Couldn't find NTLDR" 
		

When attempting to load Windows NT on a bootable array larger than 8 GBytes, or containing drive members larger than 1 GBytes capacity, NT may generate an error stating that the installation drive is greater than 1024 cylinders. NT will only allow the creation 1 GByte partitions. Therefore, the user will only be able to make a 1 GByte partition. To use a larger partition under NT:

1. Disconnect all other devices.
2. Load NT.
3. Reconnect all other devices.

During installation NT does not show the device in BOOT Order, instead it shows the array with the lower SCSI Id (on lower channel) first. To remedy :

1. Disconnect all devices other than the members of the BOOT array, so that only one disk is presented in the NT installation.
2. Configure the BOOT array in ArrayConfig such that the lowest SCSI Id on the lowest channel is the member of the BOOT array.
3. Reconnect all other devices.

When booting the NT installation CD, systems may issue the message:

		"NO-ACCESSIBLE BOOT DEVICE" 
		

1. Re-boot the NT CD-ROM.
2. At the prompt "Setup is inspecting your computers hardware configuration", press the key.
3. NT will later request the Array1000 driver diskette and the installation will succeed.

Under SCSI Adapters in NT Control Panel AAA-130 Series Array Adapters are not represented as one would expect. There will be 2, 3, or 4 icons for AAA-130 Series Array Adapter depending on 1, 2 or 3 channel card installed in the system. To check the configuration and devices connected to the AAA-130 Series Array Adapter ignore all icons representing AAA-130 Series Array adapter except the first icon. Double clicking the first icon will reveal more icons. Based on AAA-131, AAA-132 or AAA-133 Array Adapter installed there will be 1, 2 or 3 icons labeled as bus0, bus1 and bus2. Bus0, bus1 and bus2 correspond to SCSI channels 1, 2 and 3, respectively, of AAA-130 Series Array Adapters. Double clicking bus0, bus1and bus3 will reveal devices attached to the SCSI channel 1, 2, and 3, respectively. Single devices that are members of an array will not be seen as individual devices but the array itself will show up as a logical representation of the array.

Attempting to read this readme.txt file during CI/O AMS setup may result in a delay of up to 3 minutes when exiting the notepad utility. The setup utility will then continue with the setup process.

When creating a new array on a system running under Windows NT, the array is not listed as "usable" in the NT Disk Administrator until initialized. This is normal Windows NT functionality.

Do not change the system time while the CI/O AMS is running.
To change the system time one must:

1. Stop all CI/O AMS management components and RPC modules.
2. Change the system time.
3. Restart the CI/O AMS management components and RPCs.