M A B - 3 0 9 1 F C A L L E G R O 4 FUJITSU Native| Translation ------+-----+-----+----- Form 3.5"/SLIMLINE Cylinders 9491| | | Capacity form/unform 9120/11900 MB Heads 10| | | Seek time / track 8.0/ 1.0 ms Sector/track | | | Controller SCSI3 SCA/FIBRE-CHAN Precompensation Cache/Buffer 4000 KB FIFO BUFFER Landing Zone Data transfer rate 19.500 MB/S int Bytes/Sector 512 106.000 MB/S ext FIBRE Recording method PR4ML operating | non-operating -------------+-------------- Supply voltage 5 V Temperature *C 5 50 | -10 60 Power: sleep W Humidity % 20 80 | 20 80 standby W Altitude km 3.000| 12.000 idle W Shock g 5 | 60 seek W Rotation RPM 7200 read/write W Acoustic dBA spin-up W ECC Bit ON THE FLY,REED SOLOMON MTBF h 1000000 Warranty Month Lift/Lock/Park YES Certificates ********************************************************************** I N S T A L L ********************************************************************** FUJITSU MAA SERIES FIBRE CHANNEL OPERATING MANUAL Notes On Installation --------------------- Installation direction ---------------------- horizontally vertically +-----------------+ +--+ +--+ | | | +-----+ +-----+ | | | | | | | | | +-+-----------------+-+ | | | | | | +---------------------+ | | | | | | | | | | | | | | | | | | +---------------------+ | +-----+ +-----+ | +-+-----------------+-+ +--+ +--+ | | | | +-----------------+ Mounting frame structure ------------------------ To guarantee integrity of the IDD disk enclosure (DE) insulation once mounted on the frame inside the system, special attention must be given to the note below. Note: Generally, SG and FG are connected at one point in the system enclosure. Therefore, use following procedure to maintain the insulation when mounting the IDD. Use the frame with an embossed structure or the like to avoid contact between the DE base and FG. Mount the IDD with making a gap of 2.5 mm or more between the IDD and the frame of the system. The inward projection of the screw from the DE base wall at the corner must be 4 mm or less. Tightening torque of screw must be secured with 6kg-cm. Damage: To absolutely guarantee integrity of the IDD disk enclosure (DE) insulation once actually mounted to the frame inside the system, special attention must be given to the cautionary notes below. Mount the disk enclosure using a frame of the embossed structure or a similar structure providing an equivalent function and making sure there is a distance of at least 2.5 mm between this frame and the frame on the system side. The length of the portion of the mounting screw coming out of the external wall end of the mounting frame or the IDD tap surface (penetration depth) should be 4 mm or less. Limitation of side-mounting --------------------------- Mount the side using the screw holes at both the ends. Do not use the center hole. Interface connector ------------------- The connector for the fiber channel loop bus is an unshielded SCA-2 connector conforming to SCSI-3 type which has two 20-pin rows spaced 1.27 mm (0.05 inch) apart. Installation ------------ When mounting/dismounting the drive, make sure the vibration and shock requirements are met. When removing the spindle motor from the system without stopping it while power is on, pay special attention to shock in particular until the disk drive completely stops running (about 30 seconds). Mounting procedures ------------------- Since mounting the drive depends on the system cabinet structure, determine the work procedures considering the requirements specific to each system. Fix the drive in the system cabinet with four mounting screws as follows: - The drive has 8 mounting holes (both sides: 2 2, bottom: 4). Fix the drive by using four mounting holes of both sides or the bottom. - Use mounting screws whose lengths inside the drive mounting frame are 4 mm or less when the screws are tightened. - When mounting the drive, be careful not to damage parts on the PCAs. Check that the DE (signal ground) does not touch the system cabinet chassis (frame ground). There must be a 2.5 mm or more space between the DE and chassis. IMPORTANT --------- The LED lights during the IDD is executing a command. However, in same commands, the lighting time is only an instant. Therefore, it seems that the LED blinks or the LED remains off. Checking connection ------------------- When the initial operation is checked normally after power is turned on, check that the IDD is connected to the loop from the host system. Although checking the connection depends on the structure of the host system, this section describes the general procedures. Dismounting Drive ----------------- Since the method and procedure for dismounting the disk drive for replacement of the drive, etc. depends on the locker structure of the system, etc., the work procedure must be determined in consideration of the requirements specific to the system. This section describes the general procedure and notes on dismounting the drive. CAUTION 1. When dismounting the drive which is mounted on the system while power is supplied to it. The drive to be dismounted must be separated from the loop. Dismounting the drive which is not separated from the loop may cause an unexpected error. If the drive is not separated from the loop, issue an LPB to the drive from the initiator in a primitive sequence of the order set. It is recommended to stop the spindle motor prior to this loop separation operation. The spindle motor can be stopped by a START/STOP command. It takes about 30 seconds for the spindle motor to stop completely. Then, dismount the drive using the drive mounting/dismounting mechanism, etc. of the system. If the drive is dismounted while the spindle motor is running, special care is required to avoid excessive vibration or shock to the drive. It is recommended to stop the operation once the SCA connector breaks off contact and wait until the spindle motor stops (about 30 seconds) before dismount the drive. When storing or transporting the drive, put it in an antistatic bag. 2. When dismounting the drive which is mounted on the system while power is not supplied to it. Do not move the drive until the drive stops completely (about 30 seconds if the spindle motor was stopped by a START/STOP UNIT command, and about 30 seconds after powering-off when the power was simply turned off). Then, dismount the drive using the drive mounting/dismounting mechanism, etc. of the system. When storing or transporting the drive, put it in an antistatic bag. 4MB programmable multi-segment data buffer ------------------------------------------ Data is transferred between fiber-channel loop and disk media through the embedded 4MB data buffer in the IDD. This buffer can be divided into maximum 16 areas. This feature provides the suitable usage environment for users. Since the initiator can control the disconnect/reconnect timing on the fiber-channel loop by specifying the condition of stored data to the data buffer or empty condition of the data buffer, the initiator can perform the effective input/output operations with utilizing high data transfer capability of the fiber-channel regardless of actual data transfer rate of the disk drive. Read-ahead cache feature ------------------------ After executing the READ command, the IDD reads automatically and stores (prefetches) the subsequent data blocks into the data buffer (Read-ahead caching). The high speed sequential data access can be achieved by transferring the data from the data buffer without reaccessing the disk in case the subsequent command requests the prefetched data blocks. Disks ----- The disks have an outer diameter of 95 mm (3.74 inch) and inner diameter of 25 mm (0.98 inch). The disks are good for at least 10,000 contact starts and stops. Each model contains following number of disks. MAA3182:10 MAB3091/MAC3019:5 MAB3045/MAC3045:3 Heads ----- The MR (Magnet - Resistive) of the CSS (contact start/stop) type heads are in contact with the disks when the disks are not rotating, and automatically float when the rotation is started. Spindle motor ------------- The disks are rotated by a direct-drive hall-less DC motor. The motor speed is controlled by a feedback circuit using the counter electromotive current to precisely maintain the speed at 0.5% of the specified speed. Actuator -------- The actuator, which uses a rotary voice coil motor (VCM), consumes little power and generates little heat. The head assembly at the end of the actuator arm is controlled and positioned via feedback of servo information in the data. The actuator positions heads on the CCS zone over the disk and is locked by the mechanical lock when the power is off or the spindle motor is stopped. Air circulation (recirculation filter, breather filter) ------------------------------------------------------- The heads, disks, and actuator are hermetically sealed inside a disk enclosure (DE) to keep out dust and other pollutants. The DE has a closed-loop air recirculation system. Using the movement of the rotating disks, air is continuously cycled through a filter. This filter will trap any dust generated inside the enclosure and keep the air inside the DE contaminant free. To prevent negative pressure in the vicinity of the spindle when the disks begin rotating, a breather filter is attached. The breather filter also equalizes the internal air pressure with the atmospheric pressure due to surrounding temperature changes. Read/write circuit ------------------- The read/write circuit uses head LSI chips and partial response class 4 maximum likelihood (PR4ML) modulator and demodulator circuit to prevent errors caused by external noise, thus improving data reliability. Controller circuit ------------------ The controller circuit uses LSIs to increase the reliability and uses a high speed microprocessing unit (MPU) to increase the performance of the SCSI controller. The mode of connection of all devices to the fiber channel is called "node." Each node has at least one port called "N-port." In the FC-AL, a port is particularly called "node loop (NL) port." The MAAxxxxFC, MABxxxxFC and MACxxxxFC have two ports; one is connected to one FC-AL and a maximum of 126 NL ports can be connected to one loop. Loop configuration ------------------ One port incorporates a transmit circuit and receive circuit. Information is transmitted/received by differential signals via electric signal lines. This pair of signal lines is called a link. Since signals on a link are transmitted in one direction, wiring must be conducted so that it may form a loop connection on the system. In the case of the FC-AL interface, information is ansmitted/received via each node connected on a loop. Accordingly, if the power supply is off at a node connected to the loop or if interface signals cannot be transmitted/received normally, the loop will no longer function. To avoid such a nonconformity, a circuit called "port bypass circuit" is generally provided on the back plane of the system. Node addressing --------------- Each node on the fiber channel loop is assigned its specific model number (SEL ID). The SEL ID of the drive is set at the signal level on the back plane through 7 signals from SEL_0 to SEL_6 of the SCA interface connector (CN1). In signal bit weighting, SEL_6 is the MSB indicating 2 6 and SEL_0 is the LSB indicating 2 0. ********************************************************************** F E A T U R E S ********************************************************************** FUJITSU MAA FIBRE CHANNEL OPERATING MANUAL Damage ------ When mounted in the system, the disk enclosure (DE) of the IDD requires observation of the following mounting precautions to keep the DE isolated from the mounting frame. - Mount the disk enclosure using a frame of the embossed structure or a similar structure providing an equivalent function and making sure there is a distance of at least 2.5 mm between this frame and the frame on the system side. - The length of the portion of the mounting screw coming out of the external wall end of the mounting frame or the IDD tap surface (penetration depth) should be 4 mm or less. Damage to drive --------------- The hole and screwed section are for adjustment of the balance of pressure inside/outside the DE, and should not be covered with a label or seal. Seals attached to the DE are used to keep airtightness to prevent dust from entering the DE, and should not be peeled off or damaged. Data loss --------- Do not move the drive until the drive stops completely (about 30 seconds if the spindle motor was stopped by a START/STOP UNIT command, and about 30 seconds after powering-off when the power was simply turned off). Notes: When the SEND DIAGNOSTIC command terminates with the CHECK CONDITION status, the INIT must collect the error information using the REQUEST SENSE command. The RECEIVE DIAGNOSTIC RESULTS command cannot read out the error information detected in the self-diagnostics. Damage ------ Do not open the DE in the field because it is completely sealed. Data loss Save data stored on the disk drive before requesting repair. Fujitsu does not assume responsibility if data is destroyed during servicing or repair. Save data stored on the disk drive before requesting repair. Fujitsu does not assume responsibility if data is destroyed during servicing or repair. Since GaAs compound semiconductors are used in this product, it is necessary to strictly adhere to the following items to prevent danger. - Do not take them into the mouth. It is dangerous to ingest GaAs compound semiconductors used in this product. - Do not transform them to gas, powder or liquid. It is dangerous to inhale or drink by mistake the liquid generated by burning, crushing or chemically processing this product. - When disposing of this product, adhere to applicable domestic laws and your waste disposal regulations. When disposing of this product, separate it from general industrial waste and household waste. FC-AL standard -------------- The IDD provides not only FC-AL basic functions but also the following features: - Arbitration - Data frame CRC function - Command set which meets the logical specification of the SCSI CCS (Common Command Set for Direct Access Device) requirements (Rev. 4.B) The SCSI commands can manipulate data through logical block addressing regardless of the physical characteristics of the disk drive. This allows software to accommodate future expansion of system functions. Dual port support ----------------- In order to support dual ports, the IDD is equipped with two pairs of fiber channel driver/receiver. High speed data transfer ------------------------ The data transfer rate on the fiber channel loop is 106.25 MB/s maximum. In addition, the large capacity data buffer of the HDD allows the user to make full use of the high speed data transfer capability of the fiber channel loop. High speed spindle motor (MACxxxxFC) ------------------------------------ The MACxxxxFC type IDD increases the number of revolutions of the spindle motor from conventional 7,200 rpm to 10,033 rpm. Increasing the number of revolutions of the spindle motor will make it possible to improve the data transfer rate and reduce the rotation wait time. Continuous block processing --------------------------- The addressing method of data blocks is logical block address. The initiator can access data by specifying block number in a logically continuous data space without concerning the physical structure of the track or cylinder boundaries. The continuous processing up to [64K-1] blocks in a command can be achieved, and IDD can perform continuous read/write operation when processing data blocks on several tracks or cylinder. 4MB programmable multi-segment data buffer ------------------------------------------ Data is transferred between fiber-channel loop and disk media through the embedded 4MB data buffer in the IDD. This buffer can be divided into maximum 16 areas. This feature provides the suitable usage environment for users. Since the initiator can control the disconnect/reconnect timing on the fiber-channel loop by specifying the condition of stored data to the data buffer or empty condition of the data buffer, the initiator can perform the effective input/output operations with utilizing high data transfer capability of the fiber-channel regardless of actual data transfer rate of the disk drive. Read-ahead cache feature ------------------------ After executing the READ command, the IDD reads automatically and stores (prefetches) the subsequent data blocks into the data buffer (Read-ahead caching). The high speed sequential data access can be achieved by transferring the data from the data buffer without reaccessing the disk in case the subsequent command requests the prefetched data blocks. Command queuing feature ----------------------- The IDD can queue maximum 63 commands, and optimizes the issuing order of queued commands by the reordering function. This feature realizes the high speed processing. Reserve and release functions ----------------------------- The IDD can be accessed exclusively in the multi-host or multi- initiator environment by using the reserve and release functions. Enclosure service function -------------------------- The IDD supports an enclosure service interface (ESI) compliant with the SFF-8067. This interface provides the function of setting/reading enclosure service information using the SCSI-3 enclosure service command set (SES). Error recovery -------------- The IDD can try to recover from errors in fiber-channel loop or the disk drive using its powerful retry processing. If a recoverable data check occurs, error-free data can be transferred to the initiator after being corrected in the data buffer. The initiator software is released from the complicated error recover processing by these error recovery functions of the IDD. Automatic alternate block reassignment -------------------------------------- If a defective data block is detected during read, the IDD can automatically reassign its alternate data block. Programmable data block length ------------------------------ Data can be accessed in fixed-block length units. The data block length is programmable, and can at initializing with a multiple of four for the 512 to 528 bytes. Defective block slipping ------------------------ A logical data block can be reallocated in a physical sequence by slipping the defective data block at formatting. This results in high speed contiguous data block processing without a revolution delay due to defective data block. High speed positioning ---------------------- A rotary voice coil motor achieves fast positioning. Large capacity -------------- A large capacity can be obtained from 3.5-inch disk drives by dividing all cylinders into several partitions and changing the recording density on each partition (constant density recording). The disk subsystem with large capacity can be constructed in the good space efficiency. Start/Stop of spindle motor --------------------------- Using the SCSI command, the host system can start and stop the spindle motor. Diagnosis --------- The IDD has a diagnostic capability which checks internal controller functions and drive operations to facilitate testing and repair. Low power consumption --------------------- By using highly integrated LSI components, the power consumption of the IDD is very low, and this enables the unit to be used in wide range of environmental conditions. Low noise and low vibration --------------------------- Approx. 4.2 bels for the IDD. This makes it ideal for office use. The IDD has rubber vibration isolators, which minimize the transfer of vibration. Microcode downloading --------------------- The IDD implements the microcode download feature. This feature achieves easy maintainability of the IDD and function enhancing. Spindle motor ------------- The disks are rotated by a direct-drive hall-less DC motor. The motor speed is controlled by a feedback circuit using the counter electromotive current to precisely maintain the speed at 0.5% of the specified speed. Actuator -------- The actuator, which uses a rotary voice coil motor (VCM), consumes little power and generates little heat. The head assembly at the end of the actuator arm is controlled and positioned via feedback of servo information in the data. The actuator positions heads on the CCS zone over the disk and is locked by the mechanical lock when the power is off or the spindle motor is stopped. Air circulation (recirculation filter, breather filter) ------------------------------------------------------- The heads, disks, and actuator are hermetically sealed inside a disk enclosure (DE) to keep out dust and other pollutants. The DE has a closed-loop air recirculation system. Using the movement of the rotating disks, air is continuously cycled through a filter. This filter will trap any dust generated inside the enclosure and keep the air inside the DE contaminant free. To prevent negative pressure in the vicinity of the spindle when the disks begin rotating, a breather filter is attached. The breather filter also equalizes the internal air pressure with the atmospheric pressure due to surrounding temperature changes. Read/write circuit ------------------ The read/write circuit uses head LSI chips and partial response class 4 maximum likelihood (PR4ML) modulator and demodulator circuit to prevent errors caused by external noise, thus improving data reliability. Controller circuit ------------------ The controller circuit uses LSIs to increase the reliability and uses a high speed microprocessing unit (MPU) to increase the performance of the SCSI controller. The mode of connection of all devices to the fiber channel is called "node." Each node has at least one port called "N-port." In the FC-AL, a port is particularly called "node loop (NL) port." The MAAxxxxFC, MABxxxxFC and MACxxxxFC have two ports; one is connected to one FC-AL and a maximum of 126 NL ports can be connected to one loop. Loop configuration ------------------ One port incorporates a transmit circuit and receive circuit. Information is transmitted/received by differential signals via electric signal lines. This pair of signal lines is called a link. Since signals on a link are transmitted in one direction, wiring must be conducted so that it may form a loop connection on the system. In the case of the FC-AL interface, information is transmitted/ received via each node connected on a loop. Accordingly, if the power supply is off at a node connected to the loop or if interface signals cannot be transmitted/received normally, the loop will no longer function. To avoid such a nonconformity, a circuit called "port bypass circuit" is generally provided on the back plane of the system. Node addressing --------------- Each node on the fiber channel loop is assigned its specific model number (SEL ID). The SEL ID of the drive is set at the signal level on the back plane through 7 signals from SEL_0 to SEL_6 of the SCA interface connector (CN1). In signal bit weighting, SEL_6 is the MSB indicating 2 6 and SEL_0 is the LSB indicating 2 0. The SEL ID setting of the drive ranges from #0(x'00') to #125(x'7D'). Error rate ---------- Errors detected during initialization and replaced by alternate block assignments are not included in the error rate. Data blocks to be accessed should be distributed over the disk medium equally. Unrecoverable error rate ------------------------ Errors which cannot be recovered within 63 retries and ECC correction should not exceed 10 per 10 15 bits. Positioning error rate ---------------------- Positioning errors which can be recovered by one retry should be 10 or less per 10 8 seeks. Mean Time To Repair (MTTR) -------------------------- MTTR is the average time taken by a well-trained service mechanic to diagnose and repair a drive malfunction. The drive is designed for a MTTR of 30 minutes or less. Service life ------------ The service life under suitable conditions and treatment is as follows. The service life is depending on the environment temperature. Therefore, the user must design the system cabinet so that the average DE surface temperature is as possible as low. - DE surface temperature: 45*C or less 5 years - DE surface temperature: 46*C to 50*C 4 years - DE surface temperature: 51*C to 55*C 3 years - DE surface temperature: 56*C and more strengthen cooling power so that DE surface temperature is 55*C or less. Even if the IDD is used intermittently, the longest service life is 5 years. Note: The "average DE surface temperature" means the average temperature at the DE surface throughout the year when the IDD is operating. Data Space ---------- The IDD manages the entire data storage area divided into the following three data spaces. - User space: Storage area for user data - Internal test space: Reserved area for diagnostic purposes - System space: Area for exclusive use of IDD itself The user space allow a user access by specifying data. These space can be accessed with the logical data block addressing method. The internal test space is used by Read/write test of self-diagnostics test, but user can't use direct access. The system space is accessed inside the IDD at power-on or during the execution of a specific command, but the user cannot directly access the system space. Cylinder configuration ---------------------- The IDD allocates cylinders to the user space, Internal test space, and system space. Spare areas (alternate areas) for defective sectors are provided in the user space. Several sectors in the last track of one cylinder and several cylinders (alternate cylinders) in the user space are allocated as alternate areas according to the user's assignment (MODE SELECT command). Track skew and cylinder skew ---------------------------- To avoid waiting for one turn involved in head and cylinder switching, the first logical data block in each track is shifted by the number of sectors (track skew and cylinder skew) corresponding to the switching time. At the head switching location in a cylinder, the first logical data block in track t + 1 is allocated at the sector position which locates the track skew behind the sector position of the last logical data block sector in track t. At the cylinder switching location, like the head switching location, the first logical data block in a cylinder is allocated at the sector position which locates the cylinder skew behind the last logical sector position in the preceding cylinder. The last logical sector in the cylinder is allocated when formatting, and is an unused spare sector. The number of physical sectors (track skew factor and cylinder skew factor) corresponding to the skew time varies depending on the logical data block length because the track skew and the cylinder skew are managed for individual sectors. The IDD automatically determines appropriate values for the track skew factor and the cylinder skew factor according to the specified logical data block length. The value can be read out by the MODE SENSE or MODE SENSE EXTENDED command after the track has been formatted. Format capacity --------------- The size of the usable area for storing user data on the IDD (format capacity) varies according to the logical data block or the size of the spare sector area. Defect list ----------- Information of the defect location on the disk is managed by the defect list. The following are defect lists which the IDD manages. - P list (Primary defect list): This list consists of defect location information available at the disk drive shipment and is recorded in a system space. The defects in this list are permanent, so the INIT must execute the alternate block allocation using this list when initializing the disk. - D list (Data defect list): This list consists of defect location information specified in a FORMAT UNIT command by the INIT at the initialization of the disk. This information is recorded in the system space of the disk drive as the G list. To execute the alternate block allocation, the FORMAT UNIT command must be specified. - C list (Certification defect list): This list consists of location information on defective blocks which are detected by the verifying operation (certification) of the data block after the initiation when executing the FORMAT UNIT command. The IDD generates this information when executing the FORMAT UNIT command, and the alternate block allocation is made upon the defective block. This information is recorded in the system space of the disk drive as the G list. - G list (Growth defect list): This list consists of defective logical data block location information specified in a REASSIGN BLOCKS command by the INIT, information on defective logical data blocks assigned alternate blocks by means of IDD automatic alternate block allocation, information specified as the D list, and information generated as the C list. They are recorded in the system space on the disk drive. The INIT can read out the contents of the P and G lists by the READ DEFECT DATA command. Alternate block allocation -------------------------- The alternate data block is allocated to a defective data block (= sectors) in defective sector units by means of the defect management method inside the IDD. The INIT can access all logical data blocks in the user space, as long as there is no error. Spare sectors to which alternate blocks are allocated can be provided in either "spare sectors in a cylinder" or "alternate cylinders". The INIT can specify the size and area for spare sectors by the MODE SELECT command at the time of the initialization of the disk. Both of the following are applicable to the alternate block allocation. - Sector slip treatment: Defective sectors are skipped and the logical data block corresponding to those sectors is allocated to the next physical sectors. This treatment is made on the same cylinder as the defective sector's and is effective until all spare sectors in that cylinder are used up. - Alternate sector treatment: The logical data block corresponding to defective sectors is allocated to unused spare sectors in the same cylinder or unused spare sectors in the alternate cylinder. The alternate block allocation is executed by the FORMAT UNIT command, the REASSIGN BLOCKS command, or the automatic alternate block allocation. The logical data block is allocated to the next physically continued sectors after the above sector slip treatment is made. On the other hand, the logical data block is allocated to spare sectors which are not physically consecutive to the adjacent logical data blocks. If a command which processes several logical data blocks is specified, the IDD processes those blocks in ascending order of logical data block. Alternate block allocation during FORMAT UNIT command execution When the FORMAT UNIT command is specified, the allocation of the alternate block to those defective sectors included in the specified lists (P, G, or D) is continued until all spare sectors in the same cylinder are used up. When they are used up, unused spare sectors in the alternate cylinder are allocated to the subsequent sectors in the cylinder by means of alternate sector treatment. Formatting ---------- Since the disk drive is formatted with a specific (default) data format for each model (part number) when shipped from the factory, the disk need not be formatted (initialized) when it is installed in the system. However, when the system needs data attributes different from the default format, all sides of the disk must be formatted (initialized) according to the procedures below. The user can change the following data attributes at initialization: - Logical data block length - Number of logical data blocks or number of cylinders in the user space - Alternate spare area size Self-diagnostics ----------------- The IDD has the following self-diagnostic function. This function checks the basic operations of the IDD. - Initial self-diagnostics - Online self-diagnostics (SEND DIAGNOSTIC command) For a general check of the IDD including the operations of the host system and interface, use a test program that runs on the host system.