Self managing fixed configuration raid disk in headless appliance

FIELD OF THE INVENTION

[0001] The present invention relates to the field of computer entities, and particularly although not exclusively, to “headless” computer entities which lack a conventional user interface having visual display unit, keyboard and the like.

BACKGROUND TO THE INVENTION

[0002] A conventional computer entity typically comprises a casing containing a processor, memory input/output ports and the like, a video monitor, a keyboard, and a tactile device for driving a graphical user interface, e.g. a mouse, trackball device or the like. Such prior art computers are well-known.

[0003] Another type of known computer entity is a headless computer entity, also called a “headless appliance”. Headless computer entities prevent direct administration as they do not have a monitor, mouse or keyboard, to allow direct human intervention.

[0004] Headless computer entities have an advantage of relatively lower cost due to the absence of monitor, keyboard and mouse devices.

[0005] However, because headless computer entities do not have conventional user interfaces for human interaction, and generally do not have ports by which conventional keyboard or video monitors can be connected, this creates problems in manufacture and maintenance of headless appliances, particularly in loading of operating systems and software into such devices, and in making sure that the device is self-correcting and reliable, since maintenance of the device by an end customer will not be practicable, and any faults will result in a service call out from a manufacturer of the device.

[0006] Prior art headless computer entities containing RAID disk configurations require manual user configuration after the computer entity is installed and commissioned. Further, if a RAID volume fails, or a RAID card fails, then prior art headless appliances require manual reconfiguration. This involves a service call out from a manufacturer or servicing supplier.

[0007] Two main problems occur in the case of a headless computer entity having an array of disk drives, such as a RAID.

[0008] Firstly, there is the problem of how to actually manufacture the headless computer entity containing the plurality of disk drives, given that there is no console interface via which a human operator can configure the disk drives.

[0009] Secondly, having manufactured the headless computer entity, and logically configured that entity, there is the problem of maintaining the logical configuration of the RAID array in the computer entity.

[0010] In a conventional server arrangement having a user console, after manufacture, a human administrator can configure a RAID array to a required state, and then install an operating system into the RAID array, followed by application programs and data.

[0011] Additionally, in a conventional computer entity having a user console with video display unit, keyboard, pointing device and the like, a human administrator can reconfigure the logical configuration of the computer entity after installation, and attend to any error conditions which have occurred.

[0012] Neither of these options are available in a headless computer entity because there is no user console.

SUMMARY OF THE INVENTION

[0013] One object of the present invention is to provide a headless computer entity capable of self installation of an operating system which uses a fixed configuration RAID data storage device.

[0014] This may provide an advantage of easier installation of a RAID disk subsystem in the environment of a headless computer entity. A further advantage may include increased reliability by avoiding the possibility of user error in installation of the RAID disk configuration.

[0015] A second object of the present invention is to provide a fully self managing RAID disk subsystem which automatically creates and/or repairs a RAID disk configuration without any user intervention.

[0016] According to a first aspect of the present invention there is provided a method of manufacturing a headless computer entity, said entity comprising an array of a plurality of physical data storage devices, said method comprising the steps:

[0017] introducing a partition structure into a first said physical data storage device of said array of physical data storage devices;

[0018] replicating said partition structure to a second said physical data storage device of said array to create a first data volume comprising said first and second physical data storage devices;

[0019] configuring said plurality of physical data storage devices to create a second data volume, without erasing said partitioning;

[0020] wherein, said replication of said first physical data storage device, and said creation of said data volumes, is controlled by an instruction set stored on said first physical data storage device.

[0021] According to a second aspect of the present invention there is provided a headless computer entity comprising:

[0022] a redundant array device comprising a plurality of physical data storage devices;

[0023] a management module for managing said array of physical data storage devices;

[0024] said array of a plurality of physical data storage devices configured into at least one RAID 1 volume, wherein said management module is stored as program data in said RAID 1 volume.

[0025] According to a third aspect of the present invention there is provided a physical data storage device capable of attaching in a redundant array of data storage devices, said physical data storage device comprising:

[0026] program data implementing a program for managing said redundant array of data storage devices;

[0027] data implementing a driver for interfacing with said redundant array of data storage devices;

[0028] wherein said program device operates to configure said redundant array of data storage devices when said physical data storage device is installed in said array, without deleting itself.

[0029] According to a fourth aspect of the present invention there is provided a method of operating a headless computer entity, said entity comprising an array of a plurality of physical data storage devices, said method comprising the steps:

[0030] introducing a data storage device management program to a first said physical data storage device of said array; and

[0031] running said management program to create a plurality of data volumes across said plurality of physical data storage devices, without erasing said management program.

BRIEF DESCRIPTION OF THE DRAWINGS

[0032] For a better understanding of the invention and to show how the same may be carried into effect, there will now be described by way of example only, specific embodiments, methods and processes according to the present invention with reference to the accompanying drawings in which:

[0033]
FIG. 1 illustrates schematically an external view of a headless computer entity according to a specific implementation of the present invention;

[0034]
FIG. 2 illustrates schematically an internal component architecture of the headless computer entity containing hardware and logical components;

[0035]
FIG. 3 illustrates schematically an arrangement of individual disk devices in a RAID configuration within the headless computer entity of FIGS. 1 and 2;

[0036]
FIG. 4 illustrates schematically a logical partitioning of a RAID volume within the headless computer entity;

[0037]
FIG. 5 illustrates schematically a logical architecture of a RAID data storage device subsystem.

[0038]
FIG. 6 illustrates schematically control of a RAID subsystem by a RAID management application via a RAID driver within the headless computer entity;

[0039]
FIG. 7 illustrates schematically a process for initial installation of a system disk into the headless computer entity, and configuration of and creation of RAID volumes, whilst preserving system disk partitions;

[0040]
FIG. 8 illustrates schematically process steps carried out by a RAID array management application, triggered by boot of the headless computer entity;

[0041]
FIG. 9 illustrates schematically a process for querying a RAID card for configuration information, and data storage capacity;

[0042]
FIG. 10 illustrates schematically process steps carried out by a RAID management apparatus for repairing a RAID 5 configuration within a RAID data storage device subsystem;

[0043]
FIG. 11 illustrates schematically process steps carried out by a RAID management module on detecting a failure to first boot from a factory default system disk for installing an operating system and configuring a RAID data storage system;

[0044]
FIG. 12 illustrates schematically process steps carried out by the RAID management module for detecting a repair of a RAID 5 volume;

[0045]
FIG. 13 illustrates schematically process steps carried out by RAID management module in response to a RAID card failure of a RAID data storage subsystem;

[0046]
FIG. 14 illustrates schematically process steps carried out by RAID management module on detection of a single hard disk failure within the RAID subsystem;

[0047]
FIG. 15 illustrates schematically process steps carried out by a RAID management module on detection of multiple hard disk failures in a RAID 5 volume;

[0048]
FIG. 16 illustrates schematically process steps carried out by the RAID management module on detection of failure of 4 disks connected to a second or third SCSI channel; and

[0049]
FIG. 17 illustrates schematically process steps carried out by the RAID management module for RAID configuration rebuild.

DETAILED DESCRIPTION OF THE BEST MODE FOR CARRYING OUT THE INVENTION

[0050] There will now be described by way of example the best mode contemplated by the inventors for carrying out the invention. In the following description numerous specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent however, to one skilled in the art, that the present invention may be practiced without limitation to these specific details. In other instances, well known methods and structures have not been described in detail so as not to unnecessarily obscure the present invention.

[0051] Referring to FIG. 1 herein there is illustrated in perspective view, a headless computer entity 100. The headless computer entity may be configured to perform a specific computing task, for example as a network attached storage device (NAS). The headless computer entity comprises a casing 101, containing a processor, memory, data storage devices e.g. a hard disk, or a RAID array of disks, or a disk mirror arrangement and a communications port connectable to a local area network cable 102; a small display, for example a liquid crystal display (LCD) 103 giving limited information on a status of the device, for example POWER ON, STANDBY or other modes of operation; optionally, a CD ROM drive 104, and optionally back-up tape data storage device 105. Otherwise the headless computer entity has no user interface, and is self-maintaining when in operation. Direct human intervention with the headless computer entity is restricted by the lack of user interface. In operation, the headless computer entity is intended to be self-managing and self-maintaining.

[0052] Referring to FIG. 2 herein, there is illustrated schematically physical and logical components of the computer entity 100. The computer entity comprises a communications interface 201, for example a local area network card such as an Ethernet card; a data processor 202, for example an Intel® Pentium or similar Processor; a memory 203, a data storage device 204, in the best mode herein an array of individual disk drives in a RAID (redundant array of inexpensive disks) configuration; an operating system 205, for example the known Windows 2000®, Windows95, Windows98, Unix, or Linux operating systems or the like; a display 206, such as an LCD display; an administration interface 207 by means of which information describing the status of the computer entity can be communicated to a remote display; a management module 208 for managing the data storage device 204; and plurality of applications programs 209.

[0053] Within the operating system, there is included a RAID driver which communicates with the RAID management card in the RAID device hardware. The RAID driver provides an application programming interface (API) through which various instructions can be sent to the RAID management card for the RAID management card to configure the array of physical disks into RAID volumes. Typically communication between the API and the RAID management card is by proprietary protocol.

[0054] The applications programming interface allows control of the RAID management card via a RAID driver by an application program. The pre-written application program is contained on the system disk, and is configured to manage the RAID data storage device in the specific implementation described herein according to the specific processes described herein.

[0055] Since all the operating system and software is pre-installed on the headless computer entity, and there is no other way to install it, the operating system and applications, which includes the programs for setting up the RAID configuration, are installed before the RAID volumes are automatically configured.

[0056] This leads to the problem that to create a RAID 0, RAID 3, or RAID 5 volumes erases all the data on the disks, since these RAID levels stripe the data across a plurality of disks. Therefore, a RAID 1 mirror pair architecture is used for the system disk, and a more efficient RAID level, for example RAID 5 is used for the RAID data volume. The proposed RAID configuration is illustrated in FIG. 3 herein.

[0057] Referring to FIG. 3 herein, there is illustrated schematically a construction and configuration of the data storage device 204. The data storage device comprises a RAID management card 300, the RAID management card connected to a back plane 301, being a physical construction having a plurality of bays, each bay having physical locating means and electrical connectors, for connecting to an individual hard disk drive unit, such that plurality of hard disk drive units 302-313 slot into corresponding respective bays of the back plane, and receive power supply and read/write signals from the back plane, the whole unit being managed by the RAID management card 300. RAID array devices are well known in the prior art.

[0058] The RAID management card 300 effectively abstracts the plurality of physical disks, and presents these as a plurality of logical RAID volumes to the operating system 205. The operating system communicates with the RAID card, which then serves data storage capacity to the operating system, configured as logical RAID volumes of data storage capacity. The RAID card converts the plurality of physical disks, into a smaller number of logical RAID volumes.

[0059] When a newly manufactured RAID device is included in the headless computer entity, when the RAID card is initially powered up, it must be instructed on how to configure the RAID device into a plurality of RAID volumes.

[0060] Further, if there is a fault in the RAID device, either in the card, or in one or more of the physical disk devices, the RAID management card must configure the physical disk drives into a RAID volume configuration suitable for the computer entity.

[0061] In the case of new manufacture, there is a problem in how the RAID management card configures the RAID volumes without human intervention during the manufacture process. Human intervention is prohibited by the lack of user console.

[0062] In the case of failure mode, there is the problem of how to reconfigure the RAID device, minimising loss of data and preserving user data where possible, bearing in mind the absence of administrator access, due to the lack of a user console.

[0063] Referring to FIG. 4 herein, there is shown an example of an array of a plurality of 12 disks in a RAID configuration, implementing data storage device 204. Although there are 12 individual physical disks, logically the disks constitute a single data storage capability divided up into a system area, and a data area. This division needs to be set up during the manufacture process of the headless computer entity.

[0064] The RAID device can be configured into different types of RAID volume, having different levels of redundancy. In the example shown in FIG. 4, the RAID device is divided into a RAID 1 volume, and a RAID 5 volume. The RAID 1 volume comprises 2 physical disks 302, 303 whereas the RAID 5 volumes comprises the remaining 10 physical disks in the RAID device. The RAID 1 volume is assigned as a system area and the RAID 5 volume is assigned as a data storage volume and for storage of application data.

[0065] In the RAID array, all disks should be of the same capacity. Otherwise, if one disk has a smaller capacity than the remaining disks, the configuration of the RAID volumes is limited by the capacity of the smallest disk.

[0066] Optionally, default application data may be copied from the system disk onto the RAID 5 volume configured as a data partition the OSBA on the system disk contains a known good complete copy of a primary operating system, and default data sets for the primary data partition and secondary data partition. This application data may be copied from the OSBA onto the RAID 5 volume configured as a logical data storage partition. The application executable files remain in the operating system in the OSBA, whereas the application data may reside in the RAID 5 volume secondary data partition, this may include a SQL server data base for example one part of the data base may be on the logical system disk, whereas another piece of the data base may be on the logical data disk, arranged that way to enhance performance, with the highest performance requirement items being stored on the logical system disk, with the lower performance requirement items being stored on the logical data partition, because the RAID 1 configuration is faster operating for writing data than the RAID 5 configuration. This can only be done if the RAID device is configured before other applications to configure the partitions on the RAID device run.

[0067] In a headless computer entity disclosed herein, because an administrator has no direct access to the entity via a user console, maintenance of the state and condition of the headless computer entity must be managed by the operating system, and application algorithms and routines of the computer entity itself. If the fundamental logical structure of the RAID array fails, then a human operator cannot intervene via a user console to reconfigure the RAID array.

[0068] Therefore, the RAID management module 208, which may be implemented either in the operating system and/or applications, must be capable of recreating the logical RAID array under a wide range of error conditions.

[0069] Features of the self managing headless computer entity having a RAID array include:

[0070] The possibility of easily manufacturing a headless computer entity and installing an operating system and application programs before a logical RAID configuration is created.

[0071] Detection of when a RAID disk subsystem needs to be configured and selection of configuration processes depending upon the circumstances of the RAID disk subsystem. For example, if a RAID card has been replaced, a new card needs to be reconfigured to match a fixed RAID volume configuration without erasing the existing RAID volumes on the disks, containing all data on the computer entity.

[0072] Handling a RAID data volume failure, and after repair of the RAID data volume, repairing a RAID configuration, but without erasing an operating system and applications which are used to perform the RAID configuration.

[0073] In a manufacturing process for manufacturing the headless computer entity described herein, a system disk master disk comprising a known good master copy of the operating system is introduced into the data storage device 204, containing a plurality of partitions.

[0074] Management of the RAID volumes is provided as a service by RAID management module 208. The service is activated by means of a service dependency call in the operating system installation routine, to ensure that the RAID management service starts first. The RAID management service does not report that it has started operating, until all RAID volume initialization is complete. A maximum time taken to start the RAID management service routine, including any services dependent upon the RAID management service must be within a pre-set time (for example, optimally set at around 5 minutes) within which the hardware re-sets itself.

[0075] Referring to FIG. 5 herein, there is illustrated schematically a format of data storage device 204, upon which one or more operating system(s) are stored. The data storage device is partitioned into a logical data storage area 500 which is divided into a plurality of partitions and sub-partitions according to the architecture shown. A main division into a primary partition 500 and a secondary partition 501 is made. Within the primary partition are a plurality of sub partitions including a primary operating system system partition 502 (POSSP), containing a primary operating system of the computer entity; an emergency operating system partition 503 (EOSSP) containing an emergency operating system under which the computer entity operates under conditions where the primary operating system is inactive or is deactivated; an OEM partition 504; a primary operating system boot partition 505 (POSBP), from which the primary operating system is booted or rebooted; an emergency operating system boot partition 506 (EOSBP), from which the emergency operating system is booted; a primary data partition 507 (PDP) containing an SQL data base 508, and a plurality of binary large objects 509, (BLOBs); a user settings archive partition 510 (USAP); a reserved space partition 511 (RSP) typically having a capacity of the order of 4 gigabytes or more; and an operating system back up area 512 (OSBA) containing a back up copy of the primary operating system files 513. The secondary data partition 501 comprises a plurality of binary large objects 514.

[0076] The RAID array configuration must be fixed, and automatically created, as part of the manufacturing process of the headless computer entity. Logical configuration of the RAID array must be achieved post-manufacture of the physical array.

[0077] On manufacture of the headless computer entity, a system disk is inserted into a bay of the RAID array. In the example of a 12 bay RAID array having 12 bays, each capable of containing a hard disk drive unit, the RAID device initially contains 11 blank disk drive units, and 1 disk drive unit containing a system disk, partitioned as illustrated in FIG. 5 herein.

[0078] Some types of RAID volume require that the RAID configuration is set, before data is stored in the RAID volume. For example in a RAID 5 volume, data is striped across a plurality of disk drives in a RAID array. This requires that the RAID 5 volume is configured, before the data is loaded onto the RAID 5 volume, because the data must be loaded onto each of a plurality of disks in the RAID 5 volume. In this case, a RAID 5 volume must be created before a system disk partitioning can be applied to that volume. This causes a problem in manufacturing, since the RAID volume must be configured before the system disk can be loaded on, but on the other hand it is the system disk which must do the configuring.

[0079] Therefore, in the implementation disclosed, the system disk is loaded into a RAID 1 volume, which does allow copying of data onto itself before the RAID 1 volume is configured. A RAID 1 volume, comprises a pair of physical disk drives mirrored with each other to provide a logical RAID 1 volume. Of the 6 prior art levels of RAID, RAID level 1 allows a users to transfer data onto a physical disk, without configuring the physical disk into a RAID 1 volume beforehand. Configuration of a physical disk into a RAID 1 volume can be done after data has been loaded onto that physical disk, and without erasing that already loaded data. This is not true of the other 5 known RAID levels (RAID 2 to 6), since these all involve striping of data across a plurality of physical disks, which necessarily means either configuring the RAID volume before data is loaded onto that volume, or losing the data due to reconfiguration.

[0080] The manufacturing process configures the default RAID card settings to make a first disk of the RAID (system disk) into a boot on demand volume, so that on the first boot, the system boots from the first disk on the RAID array. The system disk is generated by cloning a master disk image, containing an operating system and application programs. On first boot, the RAID management module automatically initializes the RAID volumes into the fixed configuration as shown in FIG. 4 herein. The RAID controller must be capable of background initialization, so that the RAID volumes are available for use by the computer entity immediately. The application programs detect a new blank logical data disk, and then automatically creates and formats all data partitions.

[0081] Referring to FIG. 6 herein, there is illustrated schematically connectivity between a RAID management application, and a RAID data storage subsystem for control of the RAID data storage subsystem, and between the RAID application and one or more display devices. The RAID management application 600 issues commands to a RAID management card 602 via a RAID driver 601, using application procedure interface provided by the RAID driver 601. The RAID management application 600 can issue commands to the RAID management card for configuration of other RAID physical disks 603 on a back plane, and receive status information from the RAID management card 602 concerning the status of individual disks. The RAID management application 600 can generate displays for display on a liquid crystal display 605 mounted on a casing of the computer entity, and to a web administration interface 604 accessible remotely by one or more further computer entities.

[0082] The RAID management application is automatically started as part of the boot process of the operating system.

[0083] Referring to FIG. 7 herein, there is illustrated schematically process steps for converting a blank unformatted RAID device into a RAID 1 volume and one or more RAID 5 volumes as illustrated in FIG. 4 herein. In step 700, the RAID card initially is set to a JBOD (Just a Bunch Of Disks) configuration, in which no one physical disk of a plurality of disks of the RAID device is specifically formatted into a logical volume. In step 701, a physical disk drive loaded with a system disk partitioned as shown in FIG. 5 herein is loaded into a bay of the RAID device. In the example described herein, where there are 12 different bays, for 12 physical disks, during manufacture 11 bays are occupied by physical disks, and the system disk configured as shown in FIG. 5 is loaded into the first bay. In step 702, a RAID 1 volume is created by the RAID management module 208 from the partitioned disk itself, and one other physical disk. In the best mode, the disk in the next bay (bay 2) is selected to form the RAID 1 volume. At this stage, the system disk has replicated itself into a more redundant form in a RAID 1 volume. In step 604, any data contained in the secondary data partition is copied into the RAID 5 volume from the RAID 1 volume. In step 703, one or more RAID 5 volumes are created from the remaining 10 physical disks in the array of disks. A secondary data partition 501 is assigned to the RAID 5 volume, so that the arrangement shown in FIG. 4, 2 physical disks comprise a RAID 1 volume, and a remaining 10 physical disks are configured into a RAID 5 volume.

[0084] The resultant RAID array comprising a RAID 1 volume and a RAID 5 volume is partitioned as in FIG. 5, where primary data partition 500 is assigned to a RAID 1 volume, and the secondary data partition 501 is assigned to a RAID 5 volume.

[0085] The RAID management module 208 instructs the RAID management card 300 to create a RAID 1 volume from the inserted system disk, and one other disk. Implementation of creation of the RAID 1 volume is handled by known processes and apparatus contained with the known RAID management card. The RAID management card effectively synchronizes the system disk with one other physical data disk to create the RAID 1 volume. The RAID management module must specify that the system disk is to be mirrored, not a blank disk. Otherwise, the RAID management card would wipe the system disk and replicate the blank disk.

[0086] If an application running on the headless computer entity uses a database, then a performance optimization may be to place the database transaction logs and the database tables on different logical disks. In this example, the transaction logs may be placed on a RAID 1 system disk and database tables may be placed on a RAID 5 data volume. However, when the computer entity is booted for the first time, the RAID 5 data disk may be blank, since the process of creating a RAID 5 volume erases those disks. Therefore, the management module 208 needs to create a database table on the RAID 5 data volume after this data volume is automatically created and formatted. When a master disk image of the system disk is created, it must be created on an appliance that has the RAID 5 data volume already created, so that when the applications are preinstalled onto the master disk, they will create a valid default database table on the RAID 5 data disk, which can then be copied on the operating system back up area.

[0087] After the management software creates and formats the RAID 5 volume, it moves the default application SQL database tables from the system disk operating system back up area to the newly created RAID 5 data volume.

[0088] The RAID hardware configuration may be fully automated and self managing, so that human administrators never have to manually create or repair a RAID volume configuration and initialization. This means that the RAID management programs running on the computer entity need to be able to run fully unattended, with no user interface except for a hardware failure alert message interface.

[0089] The RAID management module as well as initializing the default RAID volume configuration, is also responsible for repairing any RAID volume failures, for example where two disks in a RAID volume have failed.

[0090] The following failure types can occur:

[0091] Failure type 1: failure to first boot from factory default system disk. This first failure type is detected when the RAID management module has successfully initialized the default RAID volumes, in which case it then creates a file to indicate this state. If the system boots, and this file is missing, then the RAID management module detects that the computer entity is booting from a factory default disk.

[0092] Failure type 2: RAID 5 volume failure—single disk failed.

[0093] Failure type 3: RAID 5 volume failure, two or more disks failed.

[0094] Failure type 4: RAID 1 and/or 5 volume failure—as a result of RAID management card failure.

[0095] Failure type 5: Failure of four disks connected to second or third SCSI channels.

[0096] Referring to FIG. 8 herein, there is illustrated schematically activation of the RAID management service as part of a boot of the operating system. On initial boot of the operating system on initial installation, or on re-boot of the operating system at any time. In step 800, the operating system is booted. In step 801, the RAID driver commences. In step 802, the RAID management service takes control of boot process before the boot process goes any further. In step 803, it is checked whether this is the first boot of the operating system. If the boot is the first boot of the operating system, then in step 804, a logical system of RAID volumes is created as described with reference to FIG. 4 herein. In step 805, a set of disk capacity management utilities activate. If in step 803, the boot of the operating system is not the first boot, then the process proceeds directly to step 805 to start the disk capacity management utilities.

[0097] Referring to FIG. 9 herein, there is illustrated schematically process steps carried out in step 804 for creating a logical system disk and data disks by create RAID volumes. In step 900, the RAID management card is queried by the RAID management application, via the driver, for the configuration information, and the management application in step 901 checks that the correct number of physical disks are contained in the RAID device, by receiving status information from the RAID management card. If in step 902 the incorrect number of disks are found in the RAID device, then in step 903 an error message is posted to the web administration interface and/or liquid crystal display on the casing alerting to the fact that the RAID array does not contain the correct number of disks. However if a correct number of disks is found in step 902, 904 the disk capacities are checked, again by the RAID management application querying the RAID management card via the RAID driver. If the correct sizes are not found in step 905, then in step 906 the RAID management application generates an error message which is posted to the web administration interface and/or liquid crystal display, alerting the user that the correct disk capacity is not present in the RAID device. However, if the correct disk capacity is found in step 905, then in step 907 the RAID management application proceeds to create RAID volumes.

[0098] Referring to FIG. 10 herein, there is illustrated schematically process steps carried out by the RAID management service under conditions of a RAID 5 volume failure where 2 physical disks have failed. In step 1000, the RAID management service polls a RAID management card. In step 1001, a RAID volume failure is reported, in which 2 physical disks have failed. An error message is displayed on the web administration interface 207 and on the liquid crystal display 103 on the casing of the computer entity. In step 1003, the 2 disks failed are repaired as a result of human intervention, by replacement of physical disks in the bays of the RAID data storage device containing the failed physical disks. In step 1004, the operating system is re-booted, which results, in step 1005 of monitoring of the status of the previously failed disks in the RAID array. In step 1006, the previous RAID 5 configuration is erased, resulting in loss of data, and in step 1007 a new RAID 5 configuration is created, encompassing the replacement physical disks. In step 1008, the operating system runs a data disk replacement algorithm for replacement of the secondary data partition 501, onto the newly formed RAID 5 volume.

[0099] The first failure type, of RAID 1 volume failure triggers the recovery routine of FIG. 11.

[0100] Referring to FIG. 11 herein, on detecting a failure to first boot from a factory default system disk, in step 1100, the RAID management service checks the RAID array for a correct disk configuration in step 1101. An example of a correct disk configuration where 12 equal sized hard disks are attached to 3 SCSI channels of the RAID card. If the correct disk configuration is found in step 1101, then in step 1102, a RAID 1 volume is created, and in step 1103, 1 or more RAID 5 volumes are created. In step 1104, the RAID 1 volume and RAID 5 volume(s) are initialized. Initializing the RAID 1 volume will not erase the system disk (disk 0).

[0101] Referring to FIG. 12 herein, during a boot of the computer entity, on start up of the RAID management service, if, in step 1200, there is detected that a previously failed RAID 5 volume 1200 is now operating correctly in 1201, in step 1002, the repaired RAID 5 volume is initialized. In the example of the 12 disk RAID array, where 10 disks are assigned as a RAID 5 volume, then all 10 disks working correctly would constitute a correctly operating RAID 5 volume.

[0102] Referring to FIG. 13 herein, under a RAID card replacement failure type 1300, the RAID management service updates the RAID volume configuration in the non volatile random access memory of the RAID card in step 1301. In step 1302, the RAID management service recreates a default RAID 1 volume and the default RAID 5 volumes, without loss of data on these volumes.

[0103] In the event of a non-fatal RAID hardware failure, the RAID management module can display a set of hardware failure alert messages, on the status page of the web administration interface, and on the LCD interface 103 of the computer entity itself. Hardware failure alerts display a “critical severity” message.

[0104] Referring to FIG. 14 herein, under condition of a single hard disk failure, where a single disk drive fails in either the RAID 1 or RAID 5 volumes, in step 1400 then in step 1401 an alert message is generated by the RAID management service. The alert message contains information describing which specific disk has failed, and that no data has been lost, but the computer entity is vulnerable until the failed disk is replaced. In step 1402, the RAID management service sets an LED on the failed disk to a fault condition, so that the failed disk can be easily visually identified.

[0105] Referring to FIG. 15 herein, there is illustrated schematically process steps carried out by the RAID management service when a RAID volume failure is detected. In step 1500, a multiple hard disk failure in the RAID 5 volume is detected. A multiple hard disk failure includes 2 or more hard disks constituting the RAID 5 volume. In step 1501, any individual disk failures for which alert messages may previously have been generated are cleared for the RAID 5 volume. Alert message for a RAID 1 volume failure are not cleared. In step 1502, an alert message is generated and displayed on the web interface, and/or LCD on the computer entity casing, containing the information that there is a multiple disk failure, and that there is loss of data. The alert also contains information identifying which specific disks have failed. In step 1503, the RAID management service sets individual fault LED on each of the hard disks which have failed, to enable easy identification of those hard disks by maintenance personnel.

[0106] Referring to FIG. 16, herein, there is illustrated schematically process steps carried out by the RAID management service when a RAID SCSI channel failure is detected. If the RAID management service detects a failure of all 4 disk drives connected to either a second or third SCSI channel on the RAID card, in step 1600, then this could indicate either a failure of that SCSI channel on the RAID card, or a failure of the SCSI-IDE back plane connected to that SCSI channel. In this case, any individual disk failure alert messages which may be currently displayed on the web interface or LCD on the computer entity casing are cleared in step 1601 and in step 1602, a new alert message is generated containing the information to identify which individual 4 disks have failed, and also containing the message that the failure could be due to a failure in the RAID card, or in the back plane hardware rather than those actual disk drives. In the case where the failure is in the RAID card or in the back plane, then the users data will not be lost. However, if the failure is in the 4 disks themselves, then data will be lost. The RAID management service in step 1603, sets a fault LED on each of the 4 failed disks, to enable maintenance personnel to identify those disks.

[0107] Referring to FIG. 17 herein, when a failed disk is repaired by a maintenance personnel inserting a new disk device into a hot swap bay, as indicated by an illuminated fault LED on the bay, then the RAID management service, due to an automatic ongoing poll of the RAID array from the new disk hardware in step 1700 detects the repaired hard disk in step 1701. The service proceeds to initiate a RAID volume rebuild in step 1702 by separate routine. In step 1703, the RAID management service generates a message which is displayed on the web interface and/or LCD 103, that a RAID volume rebuild has started. In step 1704, the service resets the fault LED on the disk bay of the disk which has been replaced. In step 1705 on receiving a message from the RAID volume rebuild routine, the RAID management service generates a message in step 1706, that the RAID disk rebuild is complete. This message is displayed on the web interface and/or the LCD 103.

Self managing fixed configuration raid disk in headless appliance

Information

Publication Number

Date Filed

Date Published

Inventors

Original Assignees

CPC

US Classifications

International Classifications

Abstract

Description

Claims

Priority Claims (1)