Warnings for logical-server target hosts

Abstract

An administrator selects a representation of a logical server on a display. The logical server is stored in a logical-server database. The logical-server database stores logical-servers and specifies their resource requirements and network identities. The logical servers are deployable to physical and/or virtual real servers. A logical-server manager program responds to the selection by displaying a warning symbol. The administrator interacts with the warning symbol, and the logical-server manager program responds by displaying a message specifying that the real server is unsuitable or impaired as a host for deployment of the logical server.

Description

BACKGROUND OF THE INVENTION

Herein, related art is described to facilitate understanding of the invention. Related art labeled “prior art” is admitted prior art; related art not labeled “prior art” is not admitted prior art.

In data centers and other large computer systems, an administrator is often charged with selecting a host on which to install or to which to transfer a workload. A typical data center can involve a variety of hosting technologies, including physical and virtual-machine servers. Targets can vary in suitability for hosting a given workload. Some servers may exceed the workload's requirements, others may merely meet minimum requirements. Some servers may raise issues that may have to be addressed before the workload can be run, while other servers may be incompatible with the workload. Some management software provide rating indications for potential workload targets, while de-emphasizing (dimming) or omitting problematic or incompatible severs from a display of potential targets.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of a data center employing logical servers in accordance with an embodiment of the invention.

FIG. 2 is a schematic diagram of a portion of the data center of FIG. 1 relating to a logical server assigned to a hardware server.

FIG. 3 is a schematic diagram of a portion of the data center of FIG. 1 relating to a logical server assigned to a virtual-machine server.

FIG. 4 is a flow chart of a method implemented in part by a logical-server manager of the data center of FIG. 1.

FIG. 5 is an alternative flow chart of the method of FIG. 4 emphasizing mass storage deployment.

FIG. 6 is a sequence diagram for an evaluation method segment of the method of FIG. 4.

FIG. 7 is a sequence diagram continuing from FIG. 6.

DETAILED DESCRIPTION

The present invention provides for a management interface that represents unqualified as well as qualified target hosts for a workload. Explanatory text can be associated with representations of unqualified hosts. The explanatory text can be hidden until activated by an administrator interaction with the graphical representation. For example, the explanatory text can be presented in response to hovering or clicking on a warning icon.

This interface provides an administrator greater flexibility is selecting a target host for a workload. For example, the workload may have characteristics that negate the concerns associated with a warning. In other cases, the administrator may be able to take some action, e.g., acquiring a license or installing more memory, to overcome a disqualifying factor.

A “computer” is a machine including 1) computer-readable storage media in which data and instructions for manipulating the data are or can be encoded, and 2) one or more processors for manipulating tangible analogs of data in the course of executing tangibly encoded instructions. The functionality of a computer is defined in part by programs of instructions; these programs typically include 1) an application program that defines, at least in part, a mission of the computer, and 2) an operating system that serves as an interface between the application program and the computer. Hereinafter, the terms “program”, “instructions”, “code” and “data” refer to their physical analogs as encoded in computer-readable storage media unless otherwise indicated.

A “virtual machine” is a program of instructions that serves as an interface between an operating system and a computer. From the perspective of an operating system, a virtual machine is the functional equivalent or near equivalent of a computer. A “hypervisor” is a program of instructions that serves as an interface between one or more virtual machines and a computer.

A “network” is a set of interconnected network nodes and the interconnections over which the nodes communicate with each other. The nodes can include computers, peripherals (e.g., printers, modems), and storage devices. A computer that provides services to other computers (most networked computers qualify) is referred to as a “server”.

From the perspective of remote network nodes, some virtual machines are functionally equivalent to computers and thus can be considered “servers”. Herein, the phrases “hardware server” or “HW server” and “virtual-machine server” or “VM server” are used to distinguish a computer functioning as a server from a virtual machine functioning as a server. Herein, both HW servers and VM servers are considered “real servers” because they both have a substantial tangible presence (in the form of hardware or encoded media) at a network node and perform real services to other nodes on a network.

Herein, a “logical server” is a network management object. Typically, the code defining a logical server deployed on a real server resides primarily with system management servers rather than on the real server. Each logical server has assigned to it 1) one or more network identities, and 2) resource requirements. Other data, e.g., historical data, can also be assigned to a logical server. Such workload-related data can persist in a logical-server database, even when the logical server is dormant (not deployed), thus permitting workload migration by redeployment of a logical server. A logical server can be “deployed” or “undeployed”. When deployed, logical servers are assigned to real servers, typically on a one-to-one basis. During deployment, a logical server functions as a management proxy for the real server.

In general, the real server to which a deployed logical server is assigned “adopts” the network identity or identities of the logical server. VM servers conventionally have assignable network identities (e.g., MACs and WWNs); thus, when a logical server is assigned to a VM server, the VM server adopts the logical server's network identities as its own. HW servers typically have hardwired network identities; therefore, some enabling technology is required for an HW server to adopt the network identities of a logical server assigned to it. Virtual Connect technology (available with blade systems from Hewlett-Packard Company) is such an enabling technology.

A data center AP1 includes a central management workstation 11 and managed devices including a real hardware (HW) server 13 and a real virtual-machine (VM) server 15, as shown in FIG. 1. Servers 13 and 15 are connected with central management workstation 11 by an Internet Protocol (IP) network 17. Servers 13 and 15 are also connected to mass storage 19 via a storage-array network (SAN) 21.

Central management server 11 includes processors 23, communications devices 25, and computer-readable storage media 27. Media 27 is encoded with code 30 for computer-executable program instructions and data, including a logical-server manager 31, a real-server manager 33, a logical-server database 35, and a storage database 37.

Logical-server manager 31 defines and provides for management abstraction layer populated with logical servers including logical servers 41 and 43. The logical servers are characterized by data in logical-server database 35 and can persist even when not deployed to a real server. Logical server 41 is a management object deployed on real HW server 13, while logical server 43 is deployed to real VM server 15. From a management perspective, an operating system workload 45 running on HW server 13 runs on logical server 41, while another operating system workload 47 runs on VM server 15.

A system administrator 50 manages data center AP1 by interacting with logical-server manager 31. Logical-server manager 31, in turn, translates technology-independent inputs by system administrator 50 into technology-dependent instructions to real-server manager 33. Real-server manager 33 then implements the instructions by communicating with real servers 13 and 15 and other nodes over IP network 17.

Mass storage 19 is managed separately from IP network 17 by a storage administrator 51. Mass storage 19 includes arrays of hard disks, some of which provide allocated storage 53, some constitute un-allocated storage 55, and others provide for pre-allocated storage 57. Allocated storage 53 is divided into logical units 61 and 63, each of which is assigned to a host WWN in a logical-unit table 65. The WWN for logical unit 61 belongs to server 13, which thus has exclusive access to logical unit 61. The WWN for server 15 is assigned to LUN 63, so server 15 has exclusive access to LUN 63. LUNs are not defined for un-allocated storage 55.

Pre-allocated storage 57 includes LUNs 67 and 69. These are associated with WWNs in table 65 that belong to undeployed logical servers 71 and 73. Logical server 71 is a logical server that has been created but has not been deployed. Logical server 73 was previously deployed but then decommissioned. Note that characterizing information (e.g., capacity, RAID level) for each LUN is represented on central-management workstation 11 in storage database 37.

As shown in FIG. 2, the information in logical-server database 35 for logical server 31 includes network identities 201, resource requirements 203, and historical data 205. Network identities 201 include a soft media-access-control (S-MAC) address 207 and a soft world-wide name (S-WWN) 209. Requirements 203 include compute requirements 211 (such as a number and type of processors, main memory capacity), network requirements 213 (e.g., bandwidth and redundancy), and storage requirements 215 (e.g., capacity and RAID level).

Historical data 205 can include host data 217, e.g., a list of real servers on which logical server 31 has been deployed, and utilization data 219. Utilization data can be gathered with respect to such parameters as CPU cycles, memory, disk I/O, network I/O, and power. Utilization data can include average utilization per host as well as utilization-vs-time profiles. Other embodiments take into account inclusion and exclusion rules that dictate when multiple logical servers cannot or must reside next to one another in a visualized environment provided by logical-server manager 35. The utilization data can be used in finding workloads with utilizations that vary over time in a complementary fashion to make more efficient use of available resources. Historical host data 217 can be used in selecting target hosts that are the same or similar to other hosts on which any compatibility issues were successfully addressed.

From a management perspective, logical server 31 acts as a container for workload 45. However, unlike a virtual machine, which can also act as a container for a workload, logical server 31 does not intercept and translate communications between the workload and the underlying real server. In the case of workload 45, which includes an application 221 running on an operating system 223, operating system 223 communicates directly with HW server 13 substantially as it would in the absence of logical server 31.

Generally, deployment of a logical server on a real server is an indication that the requirements of the logical server are met by the real server. In addition, the real server adopts the network identities of the logical server. For example, HW server 13 includes processors 225, media 227, and communications devices 229. Communications devices 229 include a network interface card (NIC) 231 for interfacing with IP network 17 and a host-bus adapter 233 for interfacing with SAN 21. NIC 231 has a hardwired MAC (H-MAC) address 235; HBA 233 has a hardwired WWN (H-WWN) 237. Since MAC 235 and WWN 237 are hardwired, they cannot be changed to match the corresponding network identities of logical server 31.

In some embodiments of the invention, NICs and HBAs are used that support soft proxy MACs (S-MACs) and soft proxy WWNs (S-WWNs). A HW server with such a NIC and such an HBA can be configured to adopt the network identities of a logical server. The illustrated embodiment uses a different solution. HW server 13 is a blade in a blade system providing for virtual connection technology. A virtual-connection system 239 can assume network identities 207 and 209 and provide the necessary translation to H-MAC 235 and H-WWN 237.

No such translation is required for a logical server deployed on a virtual-machine server, as virtual-machine servers conventionally have soft network identities. As shown in FIG. 3, logical server 43 includes network identities 301, resource requirements 303, and historical data 305. Network identities 301 include an S-MAC 307 and an S-WWN 309. Resource requirements 303 include compute requirements 311, storage requirements 313, and network requirements 315. Historical data 305 includes historical host data 317 and utilization statistics and profiles 319. When logical server 43 is deployed, VM server assumes S-MAC 307 and S-WWN 309. Thus, devices communicating with VM server 15 act as though they are communicating with logical server 43.

In FIG. 3, logical server 43 is shown in a layer between VM server 15 and OS workload 47, which includes an application 321 and an operating system 323. In reality, operating system 323 communicates directly with VM server 15 without intervention by logical server 43. VM server 15 communicates through a hypervisor 325 with a hardware server 327. Hardware server 327 includes processors 329, media 331, and communications devices 333. Communications devices 333 include an NIC 335 with an H-MAC 337 and an HBA 339 with an H-WWN 341. Virtual-connection system 239 (FIG. 2) can be used with HW server 327 so that a logical server 71 (FIG. 1) can be deployed there.

Logical-server manager 31 provides for a method ME1 variously represented in FIGS. 4-7 to emphasize its different aspects. As flow charted in FIG. 4, method ME1 provides alternative approaches to creating a logical server. Method segment 401 involves creating an undeployed logical server. This can be accomplished by specifying values for each required parameter for a logical server, e.g., network identities and resource requirements. Alternatively, an existing logical server can be duplicated and edited (e.g., the network identities changed) to create a new logical server.

Method segment 403, on the other hand, creates a deployed logical server from a real server by importing specifications of the real server. The result can be a logical server with the real server's network identities and with resource requirements that match the specifications of the real server. Method segment 403 is particular useful when converting a pre-existing server for use as logical server technology is implemented on an existing data center. A variation of method segment 403 can provide for an import-copy operation in which the specifications of a real server are used to generate an undeployed logical server.

Method segment 405 provides for deployment of an undeployed logical server, such as one resulting from method segment 401. Alternatively, the undeployed logical server can be a formerly deployed logical server that has been released from its former host. As part of the deployment process, the target real server adopts (e.g., using virtual-connection technology) the network identities of the logical server.

Whether deployment is a result of importation at method segment 403 or deployment of a pre-existing logical server at method segment 405, deployment involves effectively conforming network identities. In the case of importation, the logical server adopts upon creation the pre-deployment network identity or identities of the real server. In the case of deployment of a pre-existing logical server with one or more pre-existing network identities, the real server effectively adopts at least one of the pre-existing network identities of the logical server. “Effectively adopts” encompasses “adopts” and “adoption by proxy”. The latter involves a soft proxy network identity for a hardwired network identity, as provided for by virtual-connection technology or NICs and HBAs that support soft MACs and WWNs, respectively.

Method segment 407 involves operating a logical server. This in turn can involve installing an operating system, installing an application, and running the application. Method segment 409 involves tracking hosts and operational data, including utilization (statistical and time-profile) data. Tracking the host, in this case, simply means associating the identity and specification of the host real server with the deployed logical server in logical-server database 35 (FIG. 1).

During the course of operation of data center AP1 (FIG. 1), a logical server may be released (un-deployed) from its host real server at method segment 411. Release is distinct from de-activation, which involves stopping a workload but maintaining the deployment of the logical server so that the workload can be reactivated on the same real server. In a release, the host real server typically relinquishes the network identities of the logical server so that the logical server can be re-deployed elsewhere.

Note that if the logical server was “imported” from the real server and so assumed network identities formerly associated with the real server, the real server can effectively assume new network identities when the logical server is released; this permits the logical server to maintain the merged network identities and continue to contain the workload stored in mass storage and network locations associated with those network identities.

Method segment 413 involves storing an un-deployed logical server. This storing is inherent, as logical servers are stored in logical-server database 35 (FIG. 1) whether they are deployed or not. The persistence of undeployed logical servers means that they are available to be redeployed on the same or different real server at a later time.

Method segment 413 involves evaluating potential real server targets as hosts for a selected logical server. Evaluation can be required for a logical server that has not been deployed or where the availability of a previous host has been compromised (e.g., another logical server has been deployed on it). Once potential real-server hosts have been evaluated, one can be selected for deployment at method segment 417.

At method segment 419, the logical server released at method segment 411 can be redeployed on the real server selected at method segment 417. This real server may be the original host or a different host. In the latter case, the different host can be one of the same or a different technology than the original host. For example, the original host can be a VM server and the new host can be a HW server. Where such a technology change is involved, logical-server manager 31 instructs real-server manager 33 regarding the measures required to effect the transfer.

The flow chart of FIG. 5 differentiates the actions 501 of storage administrator 51 from the actions 503 of system administrator 50 in the course of method ME1. As above, system administrator 50 creates one or more undeployed logical servers at method segment 401. At method segment 505, system administrator 50 requests that storage administrator 51 create storage volumes for the world-wide names and storage requirements associated with the logical servers. The request can specify capacity, RAID (redundant array of inexpensive disks) level, WWNs from logical server definitions, operating system type, etc.

At method segment 507, storage administrator 51 creates the requested logical storage volumes. This can involve selecting a storage array, creating storage volumes, mapping to array controller ports, setting masks for HBA WWN, setting host modes, and zoning. At method segment 509, storage administrator 51 provides storage volume information on WWN, array, LUN, etc., to system administrator 50.

At method segment 511, system administrator 50 stores the provided information in storage database 37 and updates undeployed logical servers with LUN details (array, controller port, WWN, LUN) information required for accessing the assigned storage volumes. In an alternative embodiment, a system administrator requests a pool of storage volumes meeting certain specifications; once the storage volumes are provided, the system administrator creates (at method segment 511) logical servers to access the storage volumes.

Between the request at method segment 505 and the communication of storage volume information at method segment 509, there is a request latency 513. In practice, this latency can be days or weeks long. In event, the latency may be out of control of the system administrator. However, since the request can be made long before a client requests a new server, the latter's request can be met promptly.

At method segment 515, a client requests a new server. At this point, system administrator 50 can respond without having to rely further on storage administrator 51. In fact, system administrator is in the same position that would result from a request at 515 for redeployment of a formerly released logical server. In response, system administrator evaluates logical server targets at method segment 415, selects a target at method segment 417, and deploys the logical server on the selected target at method segment 419. At method segment 517, system administrator 50 installs a workload, either just the operating system or an operating system and an application. In the former case, the client may install the application.

Evaluation method segment 415 is further detailed with reference to FIGS. 6 and 7. A display 601 of central management workstation 11 (FIG. 1) is shown at time T1 showing a list 603 of logical-server operations 611-620, graphic representations 41i, 43i, 71i and 73i of logical servers 41, 43, 71, and 73, respectively, and graphic representations 13i, 15i, 16i respectively of real servers 13 and 15 (FIG. 1) and a recently added hardware server.

Import 611 relates to method segment 403 of FIG. 4. Create 612 relates to method segment 401 of FIG. 4. Duplicate 613 and edit 614 also relate to method segment 401, where a new logical server is created by duplicating an existing logical server and then editing at least the network identities. Deploy 615 relates to method segment 405, e.g., assigning a logical server to a real server. Move 616 involves an atomic combination of a release from one real server and transfer to another real server. Activate 617 involves activating a workload on a deployed logical server, while deactivating 618 involves stopping a workload on a deployed logical server. Release 619 relates to method segment 411 and involves un-deploying a deployed logical server, freeing up the host real server. Delete 620 involves removing a logical server's identity and data from logical server database 35 so that the logical server no longer exists. As indicated by cursor 621, system administrator 50 has selected deploy operation 615 at time T1.

During the deploy operation, system administrator 50 selects and begins to drag graphic representation 43i for logical server 43, as indicated by cursor 623. In response, ratings indications appear in each target representation 13i, 15i, and 16i. A “do not enter” symbol 625 indicates that server 13 is not a suitable target host for logical server 43. For example, hardware server's compute resources may not meet the specification for logical server 43; alternatively, another logical server may be deployed on server 13, leaving insufficient spare capacity for logical server 43.

To the contrary, the three dark stars 627 in representation 15i represent the highest possible rating of server 15 for deployment of logical server 43. This might be expected since server 15 previously hosted logical server 43. In the illustrated embodiment, one star indicates that the minimum resource requirements for a selected logical server are met, 2 stars indicates that recommended requirements are met, and 3 stars indicates that the recommended requirements are substantially exceeded (leaving ample headroom for peak usage times). These determinations can take into account utilization collected at method segment 409 (FIG. 4). In an alternative embodiment, a five-star rating system is used to provide finer differentiation in evaluating targets. In other embodiments, more or less fine ratings are used and different colors and symbols are used to represent the ratings.

Even a maximally rated target may not be an appropriate selection. For example, administrator 50 may know that target 15 is scheduled for downtime due to maintenance and another target is needed. Target 16i represents a newly commissioned hardware server. It has a maximal rating indicated by 3-star rating 629, but the stars of rating 629 are white instead of dark, indicating deployment involves a change of technology. Host history data collected at method segment 409 (FIG. 4) for logical server 43 indicates that logical server 43 was successfully deployed on a VM server, but has not been validated on a HW server.

In most cases, logical-server manager 31 takes care of cross-technology compatibility issues. However, the chances of compatibility issues still remain slightly higher when a cross-technology transfer is involved. Since administrator 50 may be concerned about compatibility issues, a further search for a suitable target may be in order.

Note that, at time T2, in response a default technology is indicated for each target. For server 13, the default technology is expressed as “HW server” 631y. For server 15, the default technology is “VM server” 633. For the new server, the default technology is “HW server” 635. Note that this last indication includes a drop-down menu arrow 637.

System administrator 50 can click on arrow 637 using cursor 701 as shown at time T3 in FIG. 7. The result is that default technology indication 635 expands to reveal, at time T4, a drop-down menu 703 including default technology indication 635 and an alternative technology indication “VM server” 705.

At time T5, system administrator 50 clicks on alternative technology indication 705 as indicated by cursor 707. In response, by time T6, drop-down menu 703 (see time T4) has collapsed (as indicated by arrow 637) and only alternative technology indication 705 is shown. The three white star rating 629 has been replaced by a two dark star rating 709, and a warning indicator 711 has appeared.

System administrator 50 hovers over warning indication 711 (or anywhere over representation 16i) as indicated by cursor 713 at time T7. In response, at time T8, a warning message 715 appears. For example, warning message 715 may read “A hypervisor must be installed before the selected logical server can be deployed using the selected technology.” Other warning messages may occur for other reasons. Note that a similar type of message can appear when an administrator hovers over a “do not enter” indication such as 625 in FIG. 6.

The need for a hypervisor to be installed, as indicated by message 715 at time T8, is indicated by the three white starts 629 changing to two dark stars 709 when the real-server technology changed from hardware to virtual machine. Due to overhead associated with a hypervisor, lower performance can be expected than if the logical server were assigned to a “bare metal” hardware server. System administrator 50 can now choose between better performance or a lower risk of incompatibility.

The invention provides for alternative symbols, color schemes, and degrees of precision for the ratings system. The warning messages can be displayed continuously or temporarily in response to hovering or other on-going administrator action. These and other variations upon and modifications to the illustrated embodiment are provided by the present invention, the scope of which is defined by the following claims.

Claims

1. A method comprising: a logical-server manager program interacting with hardware to concurrently display on a display representations of respective logical servers and representations of respective real servers, said real servers including at least one hardware server and at least one virtual server, said logical servers being stored in a logical-server database specifying respective resource requirements and respective network identities for said logical servers, said logical servers being deployable to physical or virtual real servers, said logical servers not being physical or virtual real servers;an administrator selecting a first representation of a first logical server for deployment on a real server;said logical-server manager program responding to the selection by representing on said display a warning symbol in association with a representation of a first real server;said administrator interacting with said warning symbol; andsaid logical-server manager program responding to said interacting by displaying a message specifying that said first real server is unsuitable or impaired as a host for deployment of said first logical server, said message indicating that said real server fails to meet resource requirements of said logical server, said resource requirements being specified in said logical-server database.

2. A method as recited in claim 1 wherein said message indicates how an impairment indicated by said warning indication can be addressed.

3. A method as recited in claim 1 wherein said message is displayed in response to a hovering over said warning indication.

4. A method as recited in claim 1 wherein said message indicates that a hypervisor must be installed on said real server so that said logical server can be deployed to said real server.

5. A system comprising non-transitory computer-readable storage media encoded with code configured to, when executed by a processor: display, on a display, representations of respective logical servers and representations of respective real servers, said real servers including at least one hardware server and at least one virtual server, said logical servers being stored in a logical-server database specifying respective resource requirements and respective network identities for said logical servers, said logical servers being deployable to physical or virtual real servers, said logical servers not being physical or virtual real servers;select and begin to drag a representation of a first representation of a first logical server on said display based on inputs from an administrator;in response to the selecting and beginning to drag, representing on said display a warning indication in association with a representation of a first real server; andin response to said administrator interacting with said warning indication, displaying a message specifying that said first real server is unsuitable or impaired as a host for deployment of said first logical server, wherein said message indicates that said real server fails to meet said resource requirements of said logical server.

6. A system as recited in claim 5 wherein said message indicates how an impairment indicated by said warning indication can be addressed.

7. A system as recited in claim 5 wherein said interacting involves hovering over said warning indication.

8. A system as recited in claim 5 wherein said message indicates that a hypervisor must be installed on said first real server so that said first logical server can be deployed to said first real server.

9. A system as recited in claim 5 further comprising said processor.