6.4 Allocation Management
The Moab Cluster ManagerTM allows global, external and internal Allocation Manager parameters to be set quickly and easily.
6.4.1 Allocation Management Overview
An allocation manager (also known as an allocation bank or cpu bank) is a software system which manages resource allocations where a resource allocation grants a job a right to use a particular amount of resources. This is not the right place for a full allocations manager overview but a brief review may point out the value in using such a system.
An allocation manager functions much like a bank in that it provides a form a currency which allows jobs to run on an HPC system. The owners of the resource (cluster/supercomputer) determine how they want the system to be used (often via an allocations committee) over a particular timeframe, often a month, quarter, or year. To enforce their decisions, they distribute allocations to various projects via various accounts and assign each account an account manager. These allocations can be for use particular machines or globally usable. They can also have activation and expiration dates associated with them. All transaction information is typically stored in a database or directory server allowing extensive statistical and allocation tracking.
Each account manager determines how the allocations are made available to individual users within his project. Allocation manager managers such as PNNL's QBank allow the account manager to dedicate portions of the overall allocation to individual users, specify some of allocations as 'shared' by all users, and hold some of the allocations in reserve for later use.
When using an allocations manager each job must be associated with an account. To accomplish this with minimal user impact, the allocation manager could be set up to handle default accounts on a per user basis. However, as is often the case, some users may be active on more than one project and thus have access to more than one account. In these situations, a mechanism, such as a job command file keyword, should be provided to allow a user to specify which account should be associated with the job.
The amount of each job's allocation charge is directly associated with the amount of resources used (i.e. processors) by that job and the amount of time it was used for. Optionally, the allocation manager can also be configured to charge accounts varying amounts based on the QOS desired by the job, the type of compute resources used, and/or the time when the resources were used (both in terms of time of day and day of week).
The allocation manager interface provides near
real-time allocation management, giving a great deal of flexibility and
control over how available compute resources are used over the medium and
long term and works hand in hand with other job management features such
as Maui's throttling policies and fairshare mechanism.
Maui's allocation manager interface(s) are defined using the AMCFG parameter. This parameter allows specification of key aspects of the interface as shown in the table below.
Configuring the allocation manager consists of two steps. The first step involves specifying where the allocation service can be found. In Maui 3.2.7 and higher, this is accomplished by setting the AMCFG parameter's SERVER attribute to the appropriate URL. In earlier releases, the HOST, PORT, and TYPE attributes must be set.
Once the interface is specified, the second step involves the scheduler to allow secure communication. As with other interfaces, this is configured using the CLIENTCFG parameter within the maui-private.cfg file as described in the Security Appendix. In the case of an allocation manager, the CSKEY and CSALGO attributes should be set to values defined during initial allocation manager build and configuration as in the example below:
# maui-private.cfg CLIENTCFG[bank] CSKEY=HMAC CSALGO=HMAC
Under this configuration, when Maui decides to start a job, it contacts the allocation manager and requests an allocation reservation, or lien be placed on the associated account. This allocation reservation is equivalent to the total amount of allocation which could be consumed by the job (based on the job's wallclock limit) and is used to prevent the possibility of allocation oversubscription. Maui then starts the job. When the job completes, Maui debits the amount of allocation actually consumed by the job from the job's account and then releases the allocation reservation or lien.
These steps transpire under the covers and should be undetectable by outside users. Only when an account has insufficient allocations to run a requested job will the presence of the allocation manager be noticed. If desired, an account may be specified which is to be used when a job's primary account is out of allocations. This account, specified using the AMCFG parameter's FALLBACKACCOUNT attribute is often associated with a low QOS privilege set and priority and often is configured to only run when no other jobs are present.
Reservations can also be configured to be chargeable. One of the big hesitations have with dedicating resources to a particular group is that if the resources are not used by that group, they go idle and are wasted. By configuration a reservation to be chargeable, sites can charge every idle cycle of the reservation to a particular project. When the reservation is in use, the consumed resources will be associated with the account of the job using the resources. When the resources are idle, the resources will be charged to the reservation's charge account. In the case of standing reservations, this account is specified using the parameter SRCFG attribute CHARGEACCOUNT. In the case of administrative reservations, this account is specified via a command line flag to the setres command.
Maui will only interface to the allocation manager when running in NORMAL mode. However, this behavior can be overridden by setting the environment variable 'MAUIAMTEST' to any value. With this variable set, Maui will attempt to interface to the allocation manager regardless of the scheduler's mode of operation.
The allocation manager interface allows a site to charge accounts in a number of different ways. Some sites may wish to charge for all jobs run through a system regardless of whether or not the job completed successfully. Sites may also want to charge based on differing usage metrics, such as walltime dedicated or processors actually utilized. Maui supports the following charge policies specified via the CHARGEPOLICY attribute:
NOTE: On systems where job wallclock limits are specified, jobs which exceed their wallclock limits and are subsequently cancelled by the scheduler or resource manager will be considered as having successfully completed as far as charging is concerned, even though the resource manager may report these jobs as having been 'removed' or 'cancelled'.
6.4.3 Allocation Manager Details
18.104.22.168 QBank Allocation ManagerQBank, developed at Pacific Northwest National Laboratory (PNNL), is a dynamic cpu bank that allows system owners and funding managers to fine tune when, where, how and to whom their resources are to be rationed. Much like a bank, but with the currency measured in computational credits instead of dollars, QBank provides an administrative interface supporting familiar operations such as deposits, withdrawals, transfers and refunds. It provides balance and usage feedback to users, managers, and system administrators. Computational resources are allocated to projects and users and full accounting is made of resource utilization.
QBank employs a debit (or credit) system in which a hold (reservation) is placed against a user's account before a job starts and a withdrawal occurs immediately after the job completes. This approach ensures requestors of a resource can only use that which has been allocated to them. Allocations for a given account can be subdivided into portions available toward different users, machines and timeframes. Presetting allocations to activate and expire in regular intervals minimizes year-end resource exhaustion and facilitates capacity planning. QBank can manage and track the use of multiple systems from a central location. Additionally, support for job charge quotes and traceback debits allows QBank to be used in meta-scheduling environments involving multiple administrative domains.
In high level summary, QBank provides the following features:
22.214.171.124 Res Allocation ManagerN/A
126.96.36.199 File Allocation ManagerN/A
188.8.131.52 Gold Allocation ManagerGold is an accounting and allocation management system being developed at PNNL under the DOE Scalable Systems Software (SSS) project. Gold is similar to QBank in that it supports a dynamic approach to allocation tracking and enforcement with reservations, quotations, etc. It offers more flexible controls for managing access to computational resources and exhibits a more powerful query interface. Gold supports hierarchical project nesting. Journaling allows the preservation of all historical state information.
One of the most powerful features is that Gold is dynamically extensible. New object/record types and their fields can be dynamically created and manipulated through the regular query language turning this system into a generalized accounting and information service. This capability is extremely powerful and can be used to provide custom accounting, meta-scheduler resource-mapping, or an external persistence interface.
Gold supports strong authentication and encryption and role based access control. A Web-accessible GUI is being developed to simplify management and use of the system. Gold will support interaction with peer accounting systems with a traceback feature enabling it to function in a meta-scheduling or grid environment. It is anticipated that a beta version of Gold will be released near 2Q04. More information about Gold can be obtained by sending email to the gold development mailing list.
|© 2001-2010 Adaptive Computing Enterprises, Inc.|