Title: System and method for fail-over switching in a disk storage medium
Abstract: A disk drive is disclosed that is configured to be accessible in fail-over via buses associated with plural host interfaces. In accordance with exemplary embodiments of the present invention, the disk drive comprises a disk storage medium, a first bus connection, and a second bus connection. The disk drive comprises a switch for selectively connecting the disk storage medium to at least one of the first and second bus connections. The disk drive also comprises an interface controller for detecting whether at least one of the first and second bus connections is active and for controlling the switch in response.
Patent Number: 6,898,730 Issued on 05/24/2005 to Hanan
| Inventors:
|
Hanan; Thomas D. (Mission Viejo, CA)
|
| Assignee:
|
Western Digital Technologies, Inc. (Lake Forest, CA)
|
| Appl. No.:
|
996686 |
| Filed:
|
November 30, 2001 |
| Current U.S. Class: |
714/7; 714/43 |
| Intern'l Class: |
G06F 011/00 |
| Field of Search: |
714/7,5,43
|
References Cited [Referenced By]
U.S. Patent Documents
Other References
Enterprise Storage Staff. ‘Serial Attached SCSI Working Group Formed’.
Nov. 26, 2001. www.internetnews.com/storage/article.php/928251.
|
Primary Examiner: Beausoliel; Robert
Assistant Examiner: Wilson; Yolanda L
Attorney, Agent or Firm: Shara, Esq.; Milad G., Burns, Doane, Swecker & Mathis LLP
Claims
1. A disk drive configured to be accessible in fail-over via buses associated
with plural host interfaces, the disk drive comprising:
a disk storage medium;
a first bus connection;
a second bus connection;
a switch for selectively connecting the disk storage medium to at least one of
the first and second bus connections; and
an interface controller for detecting whether at least one of the first and second
bus connections is active and for controlling the switch in response, wherein the
switch includes a multiplexer and the first and second bus connections and the
multiplexer are integral to the disk drive.
2. A disk drive according to claim 1, wherein the first and second bus connections
are separate initiator ports of the disk drive, the initiator ports being instantiated
by the interface controller in response to detecting which of the first and second
bus connections is active.
3. A disk drive according to claim 1, wherein the first and second bus connections
and the multiplexer are external to the disk drive.
4. A disk drive according to claim 1, wherein the multiplexer is an analog device.
5. A disk drive according to claim 1, wherein the multiplexer is a digital device.
6. A disk drive according to claim 1 wherein the multiplexer is an optical device.
7. A disk drive according to claim 1, wherein the first and second bus connections
and the multiplexer are integral to a disk drive VLSI controller chip.
8. A disk drive according to claim 1, wherein the first and second bus connections
are serial advanced technology attachment (SATA) bus connections.
9. A disk drive according to claim 1, wherein the first and second bus connections
are serial attached SCSI (SAS) bus connections.
10. A disk drive according to claim 1, wherein the first and second bus connections
are Ethernet connections.
11. A method for implementing a fail-over feature for a disk drive having an
interface controller and having access to plural host interfaces, wherein first
and second bus connections are separate initiator ports of the disk drive, the
interface controller being configured to execute the steps of:
determining when a first one of the plural host interfaces is in a failed state,
the determining comprising:
monitoring the first one of the plural host interfaces for a first signal; and
sending a second signal to a second one of the plural host interfaces in an absence
of the first signal;
initiating a selection of a second one of the plural host interfaces in response
to the step of determining; and
controlling a switch used to selectively connect a disk storage medium of the
disk drive to at least one of the first and second bus connections during the step
of initiating.
12. A method according to claim 11, comprising:
instantiating one of the initiator ports in response to detecting which of the
first and second bus connections is active.
13. A method according to claim 11, wherein the first and second bus connections
are serial advanced technology attachment (SATA) bus connections.
14. A method according to claim 11, wherein the first and second bus connections
are Ethernet connections.
15. A method according to claim 11, wherein the first signal is produced at regular
time intervals determined by the first one of the plural host interfaces.
16. A method according to claim 11, wherein the second signal is a disk-drive-initiated
interrupt signal.
Description
BACKGROUND
1. Field of the Invention
The present invention relates generally to disk drives. More particularly, the
present invention relates to a system and method for fail-over switching in a disk
storage medium.
2. Background Information
Standard communication between a host computer, such as PC-based computer,
and a hard disk drive is conventionally performed using a disk drive host interface.
The hard disk drive can be, for example, an Integrated Drive Electronics/Advanced
Technology Attachment (IDE/ATA)-compliant hard drive, a SCSI-compliant hard drive,
a fiber channel device (e.g., one or more hard drives used with associated fiber
channel switches in a multi-disk environment), or any other type of hard disk drive
or storage systems.
Many disk drive systems rely upon standardized buses to connect the disk drive
host interface to the disk drive. Generally speaking, the computer's operating
system accesses a hard disk drive as an input/output (I/O) device connected to
a bus, such as the IDE/ATA, SCSI or fiber channel bus. To communicate between the
host computer and the hard disk drive via a standard bus, information is transferred
to and from the hard disk drive via the disk drive host interface using a standard
disk drive host interface protocol. Such an information transfer vests control
with the operating system, and the hard disk drive serves as a slave to the host
computer. Command signals are supplied by the host computer to the hard disk drive
via the disk drive host interface, such that the host computer controls read and
write operations for transferring data from or to the hard disk drive.
Failures at the disk drive or at the disk drive host interface can result
in the disk drive system going offline. The resulting loss of connection can affect
system throughput performance and user application performance.
Fail-over is a backup operation that automatically and transparently switches
to a standby connection if a primary host or connection fails and redirects requests
from the failed or downed system to the backup system. For example, SCSI and fiber
channel buses support a feature that allows them to be accessed by redundant host
systems. This feature protects them from losing access to a group of disk drives
as a result of a single failure. The fail-over feature is handled by the host computer,
i.e., is host-driven.
U.S. Pat. No. 5,922,077, which is hereby incorporated by reference in its entirety,
discloses a fail-over switching system. The switching system includes a fail-over
switch in a data storage system having multiple storage device controllers separately
communicating over redundant fiber channel loops such that if a fiber channel loop
becomes disabled, the affected data storage device controller may reroute its communications
by sharing the alternate fiber channel loop.
The SCSI and fiber channel host-driven dual-initiator model involves using multiple
storage device controllers for each disk drive to maintain plural, independent,
continuously-active channels. A SCSI, host-driven, dual-initiator model for performing
fail-over is disclosed in U.S. Pat. No. 6,009,535, which is hereby incorporated
by reference in its entirety.
It would be desirable to provide a disk drive fail-over protocol using redundant
host connections.
SUMMARY OF THE INVENTION
A disk drive is disclosed that is configured to be accessible in fail-over via
buses associated with plural host interfaces. In accordance with exemplary embodiments
of the present invention, the disk drive comprises a disk storage medium, a first
bus connection, and a second bus connection. The disk drive comprises a switch
for selectively connecting the disk storage medium to at least one of the first
and second bus connections. The disk drive also comprises an interface controller
for detecting whether at least one of the first and second bus connections is active
and for controlling the switch in response.
The present invention is also directed to a method for implementing a fail-over
feature for a disk drive having an interface controller and having access to plural
host interfaces. In accordance with exemplary embodiments of the present invention,
the interface controller is configured to execute the steps of determining when
a first one of the plural host interfaces is in a failed state and initiating a
selection of a second host interface in response to the step of determining.
Exemplary embodiments are applicable to all disk drive technologies, including,
but not limited to, IDE, ATA, and serial ATA (SATA), and serial attached SCSI (SAS).
Such applicability will enhance the appeal of IDE, SATA and SAS solutions in the
higher-margin enterprise storage market.
BRIEF DESCRIPTION OF THE DRAWING FIGURES
Other objects and advantages of the present invention will become apparent
to those skilled in the art upon reading the following detailed description of
preferred embodiments, in conjunction with the accompanying drawings, wherein like
reference numerals have been used to designate like elements, and wherein:
FIG. 1A is a block diagram of a system having a disk drive that is configured
to be accessible in fail-over via buses associated with plural host interfaces,
wherein the switch and first and second bus connections are internal to the disk
drive, in accordance with an exemplary embodiment of the present invention;
FIG. 1B is a block diagram of a system having a disk drive that is configured
to be accessible in fail-over via buses associated with plural host interfaces,
wherein the switch and first and second bus connections are external to the disk
drive, in accordance with an alternate exemplary embodiment of the present invention;
FIG. 2 is a flowchart illustrating the steps carried out for implementing a
fail-over feature for a disk drive having an interface controller and having access
to plural host interfaces, in accordance with an exemplary embodiment of the present invention;
FIG. 3 is a flowchart illustrating the steps carried out for determining when
a first one of plural host interfaces is in a failed state, in accordance with
an exemplary embodiment of the present invention; and
FIG. 4 is a flowchart illustrating the steps carried out for initiating a selection
of a second host interface, in accordance with an exemplary embodiment of the present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 1A is a block diagram of a system having a disk drive
100 that is
configured to be accessible in fail-over via buses associated with plural host
interfaces, in accordance with an exemplary embodiment of the present invention.
Disk drive
100 is compatible with at least one operating system
104,
residing on at least one host computer
102, that is capable of accessing
disk drives. Disk drive
100 is also compatible with at least one disk drive
host interface
106 and
108. The host computer operating system
104
can be any operating system, such as, for example, Microsoft Windows™, Linux™,
MacOS™, Solaris™, or any other computer operating system that is
capable of accessing disk drives. The disk drive host interfaces
106 and
108 can be any host interfaces, such as, for example, an Integrated Drive
Electronics/Advanced Technology Attachment (IDE/ATA) interface, a SCSI interface,
a fiber channel interface, an Ethernet interface, an InfiniBand™ interface,
a Peripheral Component Interconnect (PCI) interface, or any other host interface
that can be used for accessing a hard disk drive.
A disk storage medium
122 of disk drive
100 can be accessed using
a head assembly
120. Disk drive
100 can include a microprocessor-based
disk controller
118 for controlling head assembly
120 for reading
from and writing to disk storage medium
122 using head assembly
120.
A hard disk controller is integrated circuit hardware and associated circuitry
that are responsible for controlling the hard disk drive. Disk controller
118
can be any part or all of the printed circuit board assembly (PCBA) of the hard
disk controller of disk drive
100. Hard disk controllers are known in the
art and described, for example, in U.S. Pat. No. 5,909,334, the disclosure of which
is hereby incorporated by reference.
According to exemplary embodiments of the present invention, disk drive
100 includes a disk storage medium, such as, for example, disk storage medium
122. Disk drive
100 also includes a first bus connection
110
and a second bus connection
112. First bus connection
110 can be
connected to disk drive host interface
106 of host computer
102.
Second bus connection
112 can be connected to disk drive host interface
108 of host computer
102. According to exemplary embodiments, the
first and second bus connections are serial advanced technology attachment (SATA)
bus connections. According to an alternate exemplary embodiment, the first and
second bus connections are serial attached SCSI (SAS) bus connections. According
to an alternate exemplary embodiment, the first and second bus connections are
Ethernet connections. However, the first and second bus connections can be any
type of bus connection that can be used to transport information between disk drive
100 and host computer
102.
Disk drive
100 includes a switch for selectively connecting the disk
storage medium to at least one of the first and second bus connections. As shown
in FIG. 1A, switch
114 is connected to first bus connection
110 and
second bus connection
112. However, switch
114 can be connected to
any number of buses. According to exemplary embodiments, switch
114 includes
a multiplexer. The multiplexer is used by switch
114 to selectively connect
to either first bus connection
110 or second bus connection
112.
According to exemplary embodiments, the multiplexer is an analog device. According
to an alternate exemplary embodiment, the multiplexer is a digital device. According
to another alternate exemplary embodiment, the multiplexer is an optical device.
Switch
114 can be any type of analog, digital, or optical multiplexer or
any other type of device that can be used to selectively connect disk storage medium
122 to one of several sources.
Disk drive
100 includes an interface controller for detecting whether
at least one of the first and second bus connections is active and for controlling
the switch in response. Interface controller
116 is connected to switch
114 to control switch
114. Interface controller
116 is also
connected to disk controller
118. Interface controller
116 can act
as a conduit for data that is passed between disk controller
116 and switch
114. Alternatively, interface controller
116 can control switch
114
without having the data pass through it between disk controller
118 and
switch
114. Interface controller
116 can be computer hardware or
firmware that resides externally to disk controller
118. Alternatively,
interface controller
116 can reside internally to disk controller
118
as, for example, hardware that forms a part of the computer circuitry of disk controller
118 or as software that resides in any form of computer memory resident
internally to disk controller
118. In addition, switch
114 can be
located either internally to or externally from interface controller
116.
According to exemplary embodiments, the first and second bus connections
are separate initiator ports of the disk drive. The initiator ports are instantiated
by the interface controller in response to detecting which of the first and second
bus connections is active. Initially, disk drive
100 is connected to, for
example, first bus connection
110. Interface controller
116 initially
instructs switch
114 to selectively connect to first bus connection
110.
However, disk drive
100 can be initially connected to any bus connection.
Interface controller
114 then monitors first bus connection
110 to
determine whether the bus connection is active.
To detect which of the bus connections is active, interface controller can use,
for example, a predetermined timeout to determine if no activity has been detected
on a particular bus connection for longer than a certain time interval. For example,
a heartbeat message can be transmitted by disk drive host interface
106
over first bus connection
110 to interface controller
116 while disk
drive host interface
106 is selectively connected to disk drive
100.
However, the timeout can use any form of message or command that indicates that
a bus connection is active. So long as interface controller
116 receives
the heartbeat message from disk drive host interface
106 over first bus
connection
110 (i.e., a timeout does not occur), disk drive
100 will
remain selectively connected to first bus connection
110.
However, if the heartbeat message from disk drive host interface
106
has ceased or otherwise stopped, then a timeout will occur. In response, interface
controller
116 can send an interrupt or any other form of "ping" message
or command to disk drive host interface
108 over first bus connection
112
to determine if second bus connection
110 is active. If interface controller
116 does not receive a response from disk drive host interface
108
within a certain timeout period, first bus connection
112 can be considered
to be inactive as well. Otherwise, interface controller
116 can control
switch
114 to selectively connect to second bus connection
112.
Upon selectively connecting to second bus connection
112, disk drive
100 can instantiate the bus connection with disk drive host interface
108
using the disk drive host interface protocol-specific initialization. Since disk
drive host interface
108 will now "see" a disk drive (i.e., disk drive
100)
that it did not detect previously, the disk drive host interfaces require the ability
to establish, during their operation, a connection with a disk drive to which it
they were not previously connected. If the disk drive host interfaces do not support
such a "hot-plugging" capability, the disk drive host interfaces can be modified
by adding either hardware or software that allows the disk drive host interfaces
to instantiate a connection during their operation, even after start-up, with a
disk drive that suddenly comes online (from the perspective of the subsequent,
selectively-connected bus connection).
According to exemplary embodiments, a plurality of disk drives
100
can be connected to host computer
102. Each of the plurality of disk drives
100 can be connected to either first bus connection
110 or second
bus connection
112. For example, all disk drives
100 can be connected
to the same bus connection at the same time. For example, all the disk drives
100
can be connected to first bus connection
110. According to exemplary embodiments,
all of the disk drives
100 can communicate with each other using, for example,
interface controller
116. The disk drives
100 can communicate with
each other over first bus connection
110 or any other bus or internal connection
in host computer
102. In the event that one of the disk drives
100
detects a failure on first bus connection
106, the disk drive
100
that detected the failure can notify all of the other disk drives
100 to
selectively connect to second bus connection
112. Alternatively, each disk
drive
100 can detect the failure independently and selectively connect to
second bus connection
112 without communicating that information to any
of the other disk drives
100.
According to exemplary embodiments, the first and second bus connections
and the multiplexer are internal to the disk drive. Thus, as shown in FIG. 1A,
first bus connection
10, second bus connection
112, and switch
114
are located internally to disk drive
100 (e.g., resident on a disk drive
VLSI controller chip). For example, disk controller
118 can be configured
as a disk drive VLSI controller chip. First bus connection
110, second bus
connection
112, and switch
114 can be configured to reside on the
disk drive VLSI controller chip. According to an alternate exemplary embodiment,
the first and second bus connections and the multiplexer are external to the disk
drive. Thus, as shown in FIG. 1B, first bus connection
110, second bus connection
112, and switch
114 are located externally from disk drive
100.
In such an alternate exemplary embodiment, the external switch
114 can be
connected to disk drive
100 through interface controller
116.
FIG. 2 is a flowchart illustrating the steps carried out for implementing a
fail-over feature for a disk drive having an interface controller and having access
to plural host interfaces, in accordance with an exemplary embodiment of the present
invention. In step
202, a determination is made as to when a first one of
the plural host interfaces is in a failed state. According to exemplary embodiments,
the first and second bus connections are serial advanced technology attachment
(SATA) bus connections. According to an alternate exemplary embodiment, the first
and second bus connections are Ethernet connections. However, the first and second
bus connections can be any type of bus connection that can be used to transport
information between a disk drive and the disk drive host interfaces of a host computer.
FIG. 3 is a flowchart illustrating the steps carried out for determining when
a first one of plural host interfaces is in a failed state, in accordance with
an exemplary embodiment of the present invention. In step
302, the first
one of the plural host interfaces is monitored for a first signal. According to
exemplary embodiments, the first signal is produced at regular time intervals determined
by the first one of the plural host interfaces. For example, the first signal can
be a heartbeat message or any other form of timeout mechanism. Thus, interface
controller
116 can monitor first bus connection
110 for a heartbeat
message from disk drive host interface
106. So long as the heartbeat message
is received by interface controller
116 at regular time intervals (i.e.,
a timeout does not occur), first bus connection
110 can be considered active.
If a timeout does occur, in step
304, a second signal is sent to a second
one of the plural host interfaces in the absence of the first signal. According
to exemplary embodiments, the second signal is a disk-rive-initiated interrupt
signal. For example, interface controller
116 can send an interrupt or any
other form of message or command to host interface
108 to determine if second
bus connection
110 is still active. If interface controller
116 does
not receive a response in a certain period of time (i.e., a timeout occurs), second
bus connection
110 can be considered inactive. Otherwise, switch
114
can selectively connect to second bus connection
112.
In step
204 of FIG. 2, a selection of a second one of the plural host
interfaces
is initiated in response to the step of determining. Thus, if first bus connection
110 is determined to be inactive, interface controller
116 can control
switch
114 to selectively connect to disk drive host interface
112.
FIG. 4 is a flowchart illustrating the steps carried out for initiating a selection
of a second host interface, in accordance with an exemplary embodiment of the present invention.
In step
402, a switch is controlled to selectively connect a disk storage
medium of the disk drive to at least one of the first and second bus connections
during the step of initiating. The disk drive accesses the plural host interfaces
using the first and second bus connections. However, the switch can be connected
to any number of bus connections and, hence, the disk drive can have access to
any number of plural host interfaces. According to exemplary embodiments, if a
timeout occurs on first bus connection
110, interface controller
116
can control switch
114 to selectively connect disk storage medium
122
to second bus connection
112 in the event that such a failure of first bus
connection
110 occurs. In step
404, one of the initiator ports is
instantiated in response to detecting which of the first and second bus connections
is active. Consequently, once switch
114 selectively connects to second
bus connection
112, second bus connection
112 is instantiated between
disk drive
100 and disk drive host interface
108.
Disk drive
100 can instantiate the connection with disk drive host interface
108 using the disk drive host interface protocol-specific initialization.
Since disk drive host interface
108 will now "see" a disk drive (i.e., disk
drive
100) that it did not see previously, the disk drive host interfaces
require the ability to establish, during their operation, a connection with a disk
drive to which it they were not previously connected. If the disk drive host interfaces
do not support such a "hot-plugging" capability, the disk drive host interfaces
can be modified by adding either hardware or software that allows the disk drive
host interfaces to instantiate a connection during their operation, even after
start-up, with a disk drive that suddenly comes online (from the perspective of
the subsequent bus connection).
In the event that second bus connection
112 fails, those of ordinary skill
in the art will recognize that a similar process can be performed to selectively
connect disk drive
100 to first bus connection
10, or any other bus
to which disk drive
100 can be connected.
*
