BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1A and 1B are front and rear perspective views, respectively, of a blade chassis in which the present invention may be implemented;
FIG. 2 is a perspective view of a disk enclosure blade which may be inserted into the chassis of FIGS. 1A and 1B;
FIG. 3 is a cut-away view of a multi-drive tray which may be inserted into the disk enclosure blade of FIG. 2;
FIG. 4 schematically illustrates power domains in a blade storage subsystem;
FIG. 5 is a more detailed block diagram of the power domains of FIG. 4 within a blade storage subsystem;
FIG. 6 illustrates the power distribution within one drive enclosure blade; and
FIG. 7 is a block diagram of a blade storage subsystem in which the present invention may be implemented.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
FIGS. 1A and 1B are front and rear perspective views, respectively, of a blade chassis 100 in which the present invention may be implemented. The chassis 100 includes a housing 102 a mid- or back-plane 104 and slots 106 into which blades, such as a drive enclosure blade (DEB) 200, are inserted from the front (FIG. 1A) to mate with appropriate connectors on the front of the mid-plane 104. The IBM eServer# BladeCenter chassis includes fourteen such slots in accessible from the front. The rear of the chassis 100 (FIG. 1B) is configured to hold additional components or modules. Such modules may include, for example, two blowers 108A, 108B, up to two redundant pairs of power supply units (PSUs) 110A, 110B, 112A, 112B, a redundant pair of serial attached SCSI (SAS) switches 114A, 114B, and a management module 116. Such components are inserted from the rear of the chassis 100 to mate with appropriate connectors on the rear of the mid-plane 104.
FIG. 2 is a perspective view of a DEB 200 which may be inserted into the chassis 100. Each DEB 200 fits into three contiguous slots 106 in the chassis 100 and up to four DEBs 200 may be installed in the chassis 100. In addition, a redundant pair of RAID controller blades (RCBs) 118A, 118B may be installed in the chassis 100. Up to eight multi-drive trays 300 may be inserted into slots in the DEB 200 along with a redundant pair of local drive controller cards 202A, 2028. The multi-drive trays 300 and controller cards 202A, 202B mate with appropriate connectors on a back-plane 204 within the DEB 200. As illustrated in the cut-away view of FIG. 3, a multi-drive tray 300 may house up to three hard disk drives (HDDs) 302A, 302B, 302C. Thus, each DEB 200 may house up to twenty-four HDDs and a full chassis 100 may house up to ninety-six HDDs.
FIG. 4 schematically illustrates subsystem power domains in the blade storage subsystem. A first pair of redundant power supply units, PSU1110A and PSU2110B, comprise a first subsystem power domain 402 supplying power to slots 1-7 in the chassis 100. A second pair of redundant power supply units, PSU3112A and PSU41128 comprise a second subsystem power domain 404 supplying power to slots 8-14. If one of the PSUs in a domain fails, service will be continued by the other PSU, thereby ensuring uninterrupted operation. In the illustrated configuration, DEB1 and DEB2200A, 200B, are wholly within the first subsystem power domain 402 and DEB4200D and the two RCBs 116A, 116B are wholly within the second subsystem power domain 404. DEB3200C, in slots 7-9, spans both subsystem power domains 402, 404.
FIG. 5 is a more detailed block diagram of the subsystem power domains 402, 404. As previously described each PSU 110A, 110B, 112A, 112B connects to the rear of the mid-plane 104 while the DEBs 200A-200D connect to the front of the mid-plane 104. The mid-plane 104 includes two pairs of parallel power buses, one pair for each subsystem power domain 402, 404. PSU1110A is coupled to a first power bus 500A and PSU2110B is coupled to a second power bus 500B and PSU3112A is coupled to a third power bus 502A and PSU4112B is coupled to a fourth power bus 502B. In the front slots 106, each DEB 200 includes four power connectors with which to couple to the mid-plane 104. In DEB1200A, the first two power connectors 1A, 1B are coupled to PSU1110A and PSU2110B, respectively and are part of a first local power domain (within the DES). Similarly, the last two power connectors 3A, 3B are coupled to PSU1110A and PSU2110B, respectively, and are part of a second local power domain. The middle two power connectors 2A, 28 are not used. DEB2200B is coupled to the first and second power buses 500A, 500B in the same manner, In DEB4200D, the first two power connectors 10A, 10B are coupled to PSU3112A and PSU41128, respectively, and are part of a first local power domain. Similarly, the last two power connectors 12A, 128 are coupled to PSU3112A and PSU4112B, respectively, and are part of a second local power domain. DEB3200C spans the two subsystem power domains 402, 404; the first two power connectors 7A, 7B are coupled to PSU1110A and PSU2110B, respectively, and are part of a first local power domain while the last two power connectors 9A, 9B are coupled to PSU3112A and PSU4112B, respectively, and are part of a second local power domain. The two RCBs 118A, 118B in chassis slots 13 and 14 are within the second subsystem power domain 404 and are each coupled to power buses 502A, 502B. RCB1118A is coupled through power connectors 13A and 138 and RCB2118B is coupled through power connectors 14A and 148. It will be appreciated that the illustrated configuration is only one example and that the present invention contemplates other configurations.
FIG. 6 illustrates the power distribution within one DEB, such as DEB1200A. Four of the multi-drive trays 300A-300D and one local drive controller card 202A are within a first local power domain 600A and the other four multi-drive trays 300E-300H and the other local drive controller card 202B are within a second local power domain 600B. Although both local power domains 600A, 600B in DEB1200A are part of the first subsystem power domain 402, in DEB3200C, the first local power domain 600A would be part of the first subsystem power domain 402 and the second local power domain 600B would be part of the second subsystem power domain 404.
FIG. 7 is a block diagram of a blade storage subsystem in which the present invention may be implemented. In addition to the previously described components, the blade storage subsystem includes redundant subsystem SCSI enclosure services (SES) modules 700A, 700B (collectively referred to hereinafter as subsystem SES module 700) within the two SAS switches 114A, 114B and a local SES module 710A, 710B within each local drive controller card 202A, 202B, respectively (and collectively referred to hereinafter as local SES module 702). The subsystem SES modules 700A, 700B and the local SES modules 710A, 710B include logic for managing the power-on of multiple HDDs in the storage subsystem.
In operation, when the subsystem is powered on, such as with a power switch on the chassis 100, the management module 116 transfers control of the power-on sequence to the subsystem SES module 700. The subsystem SES module 700 performs a discovery operation to determine how many HDDs are installed and where each is located. The location includes the location of the multi-tray module in which each HDD is installed and the location of the DEB in which the multi-tray module is installed. The location also includes the power domain in which each HDD is located. The location information is captured in a table 702 or other comparable data structure within the subsystem SES 700. Such a table may be generated the first time the subsystem is powered on and updated each time a module is inserted or removed from the chassis 100. Alternatively, the table may be generated during each power-on sequence. During the discovery operation, each local SES 710 reports the mapping of SAS port addresses to physical addresses within its DEB. The subsystem SES 700 then compiles the mapping information from the local SES modules 710 into the table 702 along with information about power domain boundaries.
The subsystem SES 700 then directs the local SES modules 710 to commence powering on the HDDs in such a way that the inrush current does not exceed the limits of any power domain. In one such sequence, the subsystem SES 700 directs specific DEBs to power-on specific HDDs in a predefined order, again established such that the inrush current does not exceed the limits of any power domain. This procedure may be particularly beneficial when a DES spans two power domains. In an alternate sequence, the subsystem SES 700 directs one local SES module 710 in each power domain to power-on the HDDs in the respective DEBs. When those two local SES modules 710 report back that the HDDs are powered on, the subsystem SES 700 directs another local SES module 710 in each power domain to power-on the HDDs in the respective DEBs. The process continues until all HDDs are powered on. In a variation of the latter process, depending upon the power domain configuration and current limitations, the subsystem SES 700 may direct more than one local SES module 710 in each power domain to power-on the HDDs. For example, in a two domain system illustrated in the Figs., powering-on the HDDs in two DEBs at the same time in the same power domain may exceed the power limits of a domain. However, the subsystem SES 700 may instead direct DEB1 and DEB4200A, 200D, in power domains 1 and 2402, 404, and DEB3200C, spanning the two power domains 402, 404, to power-on the respective HDDs.
In addition, each local SES module 710 may power-on fewer than all of the HDDs at a time in a DEB 200 if powering on all would exceed the power limits of the domain. In an alternative sequence powering-on of DEBs may be partially overlapped to speed the entire process. Once the initial power spike of one DEB has dissipated, the next DEB may be powered-on with little risk of exceeding power restrictions.
The present invention also accommodates the process of hot-plugging one or more DEBs or drive trays. It will be appreciated that hot-plugging a module can generate the same power surge that a convention power-on can generate. Consequently, in response to a signal that one or more DEBs or drive trays have been hot-plugged, the subsystem SES module 700 directs the appropriate local SES module 710 to power on the new DEBs or drives in such a manner that the power limits are not exceeded.
It is important to note that while the present invention has been described in the context of a fully functioning data processing system, those of ordinary skill in the art will appreciate that the processes of the present invention are capable of being distributed in the form of a computer readable medium of instructions and a variety of forms and that the present invention applies regardless of the particular type of signal bearing media actually used to carry out the distribution. Examples of computer readable media include recordable-type media such as a floppy disk, a hard disk drive, a RAM, and COD-ROMs and transmission-type media such as digital and analog communication links.
The description of the present invention has been presented for purposes of illustration and description, but is not intended to be exhaustive or limited to the invention in the form disclosed. It will be appreciated that the present invention is not limited to use with a subsystem of the foregoing description. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiment was chosen and described in order to best explain the principles of the invention, the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated. Moreover, although described above with respect to methods and systems, the need in the art may also be met with a computer program product containing instructions for managing the power-on of multiple hard disk drives in a storage subsystem.
Patent Application
20080028238
