Exclusive cpusets¶

A cpuset that is mem_exclusive restricts kernel allocations for buffer cache pages and other internal kernel data pages commonly shared by the kernel across multiple users. All cpusets, whether mem_exclusive or not, restrict allocations of memory for user space. This enables configuring a system so that several independent jobs can share common kernel data, while isolating each job's user allocation in its own cpuset. To do this, construct a large mem_exclusive cpuset to hold all the jobs, and construct child, non-mem_exclusive cpusets for each individual job. Only a small amount of kernel memory, such as requests from interrupt handlers, is allowed to be placed on memory nodes outside even a mem_exclusive cpuset.

Hardwall¶

This enables configuring a system so that several independent jobs can share common kernel data, such as filesystem pages, while isolating each job's user allocation in its own cpuset. To do this, construct a large hardwall cpuset to hold all the jobs, and construct child cpusets for each individual job which are not hardwall cpusets.

Only a small amount of kernel memory, such as requests from interrupt handlers, is allowed to be taken outside even a hardwall cpuset.

Notify on release¶

The default value of notify_on_release in the root cpuset at system boot is disabled (0). The default value of other cpusets at creation is the current value of their parent's notify_on_release setting.

The command /sbin/cpuset_release_agent is invoked, with the name (/dev/cpuset relative path) of the to-be-released cpuset in argv[1].

The usual contents of the command /sbin/cpuset_release_agent is simply the shell script:

#!/bin/sh
rmdir /dev/cpuset/$1

As with other flag values below, this flag can be changed by writing an ASCII number 0 or 1 (with optional trailing newline) into the file, to clear or set the flag, respectively.

Memory pressure¶

This enables batch managers that are monitoring jobs running in dedicated cpusets to efficiently detect what level of memory pressure that job is causing.

This is useful both on tightly managed systems running a wide mix of submitted jobs, which may choose to terminate or reprioritize jobs that are trying to use more memory than allowed on the nodes assigned them, and with tightly coupled, long-running, massively parallel scientific computing jobs that will dramatically fail to meet required performance goals if they start to use more memory than allowed to them.

This mechanism provides a very economical way for the batch manager to monitor a cpuset for signs of memory pressure. It's up to the batch manager or other user code to decide what action to take if it detects signs of memory pressure.

Unless memory pressure calculation is enabled by setting the pseudo-file /dev/cpuset/cpuset.memory_pressure_enabled, it is not computed for any cpuset, and reads from any memory_pressure always return zero, as represented by the ASCII string "0\n". See the WARNINGS section, below.

A per-cpuset, running average is employed for the following reasons:

• Because this meter is per-cpuset rather than per-process or per virtual memory region, the system load imposed by a batch scheduler monitoring this metric is sharply reduced on large systems, because a scan of the tasklist can be avoided on each set of queries.
• Because this meter is a running average rather than an accumulating counter, a batch scheduler can detect memory pressure with a single read, instead of having to read and accumulate results for a period of time.
• Because this meter is per-cpuset rather than per-process, the batch scheduler can obtain the key information—memory pressure in a cpuset—with a single read, rather than having to query and accumulate results over all the (dynamically changing) set of processes in the cpuset.

The memory_pressure of a cpuset is calculated using a per-cpuset simple digital filter that is kept within the kernel. For each cpuset, this filter tracks the recent rate at which processes attached to that cpuset enter the kernel direct reclaim code.

The kernel direct reclaim code is entered whenever a process has to satisfy a memory page request by first finding some other page to repurpose, due to lack of any readily available already free pages. Dirty filesystem pages are repurposed by first writing them to disk. Unmodified filesystem buffer pages are repurposed by simply dropping them, though if that page is needed again, it will have to be reread from disk.

The cpuset.memory_pressure file provides an integer number representing the recent (half-life of 10 seconds) rate of entries to the direct reclaim code caused by any process in the cpuset, in units of reclaims attempted per second, times 1000.

Memory spread¶

If the per-cpuset Boolean flag file cpuset.memory_spread_page is set, then the kernel will spread the filesystem buffers (page cache) evenly over all the nodes that the faulting process is allowed to use, instead of preferring to put those pages on the node where the process is running.

If the per-cpuset Boolean flag file cpuset.memory_spread_slab is set, then the kernel will spread some filesystem-related slab caches, such as those for inodes and directory entries, evenly over all the nodes that the faulting process is allowed to use, instead of preferring to put those pages on the node where the process is running.

The setting of these flags does not affect the data segment (see brk(2)) or stack segment pages of a process.

By default, both kinds of memory spreading are off and the kernel prefers to allocate memory pages on the node local to where the requesting process is running. If that node is not allowed by the process's NUMA memory policy or cpuset configuration or if there are insufficient free memory pages on that node, then the kernel looks for the nearest node that is allowed and has sufficient free memory.

When new cpusets are created, they inherit the memory spread settings of their parent.

Setting memory spreading causes allocations for the affected page or slab caches to ignore the process's NUMA memory policy and be spread instead. However, the effect of these changes in memory placement caused by cpuset-specified memory spreading is hidden from the mbind(2) or set_mempolicy(2) calls. These two NUMA memory policy calls always appear to behave as if no cpuset-specified memory spreading is in effect, even if it is. If cpuset memory spreading is subsequently turned off, the NUMA memory policy most recently specified by these calls is automatically reapplied.

Both cpuset.memory_spread_page and cpuset.memory_spread_slab are Boolean flag files. By default, they contain "0", meaning that the feature is off for that cpuset. If a "1" is written to that file, that turns the named feature on.

Cpuset-specified memory spreading behaves similarly to what is known (in other contexts) as round-robin or interleave memory placement.

Cpuset-specified memory spreading can provide substantial performance improvements for jobs that:

a)

need to place thread-local data on memory nodes close to the CPUs which are running the threads that most frequently access that data; but also

b)

need to access large filesystem data sets that must to be spread across the several nodes in the job's cpuset in order to fit.

Without this policy, the memory allocation across the nodes in the job's cpuset can become very uneven, especially for jobs that might have just a single thread initializing or reading in the data set.

Memory migration¶

When memory migration is enabled in a cpuset, if the mems setting of the cpuset is changed, then any memory page in use by any process in the cpuset that is on a memory node that is no longer allowed will be migrated to a memory node that is allowed.

Furthermore, if a process is moved into a cpuset with memory_migrate enabled, any memory pages it uses that were on memory nodes allowed in its previous cpuset, but which are not allowed in its new cpuset, will be migrated to a memory node allowed in the new cpuset.

The relative placement of a migrated page within the cpuset is preserved during these migration operations if possible. For example, if the page was on the second valid node of the prior cpuset, then the page will be placed on the second valid node of the new cpuset, if possible.

Scheduler load balancing¶

The algorithmic cost of load balancing and its impact on key shared kernel data structures such as the process list increases more than linearly with the number of CPUs being balanced. For example, it costs more to load balance across one large set of CPUs than it does to balance across two smaller sets of CPUs, each of half the size of the larger set. (The precise relationship between the number of CPUs being balanced and the cost of load balancing depends on implementation details of the kernel process scheduler, which is subject to change over time, as improved kernel scheduler algorithms are implemented.)

The per-cpuset flag sched_load_balance provides a mechanism to suppress this automatic scheduler load balancing in cases where it is not needed and suppressing it would have worthwhile performance benefits.

By default, load balancing is done across all CPUs, except those marked isolated using the kernel boot time "isolcpus=" argument. (See Scheduler Relax Domain Level, below, to change this default.)

This default load balancing across all CPUs is not well suited to the following two situations:

• On large systems, load balancing across many CPUs is expensive. If the system is managed using cpusets to place independent jobs on separate sets of CPUs, full load balancing is unnecessary.
• Systems supporting real-time on some CPUs need to minimize system overhead on those CPUs, including avoiding process load balancing if that is not needed.

When the per-cpuset flag sched_load_balance is enabled (the default setting), it requests load balancing across all the CPUs in that cpuset's allowed CPUs, ensuring that load balancing can move a process (not otherwise pinned, as by sched_setaffinity(2)) from any CPU in that cpuset to any other.

When the per-cpuset flag sched_load_balance is disabled, then the scheduler will avoid load balancing across the CPUs in that cpuset, except in so far as is necessary because some overlapping cpuset has sched_load_balance enabled.

So, for example, if the top cpuset has the flag sched_load_balance enabled, then the scheduler will load balance across all CPUs, and the setting of the sched_load_balance flag in other cpusets has no effect, as we're already fully load balancing.

Therefore in the above two situations, the flag sched_load_balance should be disabled in the top cpuset, and only some of the smaller, child cpusets would have this flag enabled.

When doing this, you don't usually want to leave any unpinned processes in the top cpuset that might use nontrivial amounts of CPU, as such processes may be artificially constrained to some subset of CPUs, depending on the particulars of this flag setting in descendant cpusets. Even if such a process could use spare CPU cycles in some other CPUs, the kernel scheduler might not consider the possibility of load balancing that process to the underused CPU.

Of course, processes pinned to a particular CPU can be left in a cpuset that disables sched_load_balance as those processes aren't going anywhere else anyway.

Scheduler relax domain level¶

On small systems, such as those with just a few CPUs, immediate load balancing is useful to improve system interactivity and to minimize wasteful idle CPU cycles. But on large systems, attempting immediate load balancing across a large number of CPUs can be more costly than it is worth, depending on the particular performance characteristics of the job mix and the hardware.

The exact meaning of the small integer values of sched_relax_domain_level will depend on internal implementation details of the kernel scheduler code and on the non-uniform architecture of the hardware. Both of these will evolve over time and vary by system architecture and kernel version.

As of this writing, when this capability was introduced in Linux 2.6.26, on certain popular architectures, the positive values of sched_relax_domain_level have the following meanings.

(1)

Perform immediate load balancing across Hyper-Thread siblings on the same core.

(2)

Perform immediate load balancing across other cores in the same package.

(3)

Perform immediate load balancing across other CPUs on the same node or blade.

(4)

Perform immediate load balancing across over several (implementation detail) nodes [On NUMA systems].

(5)

Perform immediate load balancing across over all CPUs in system [On NUMA systems].

The sched_relax_domain_level value of zero (0) always means don't perform immediate load balancing, hence that load balancing is done only periodically, not immediately when a CPU becomes available or another task becomes runnable.

The sched_relax_domain_level value of minus one (-1) always means use the system default value. The system default value can vary by architecture and kernel version. This system default value can be changed by kernel boot-time "relax_domain_level=" argument.

In the case of multiple overlapping cpusets which have conflicting sched_relax_domain_level values, then the highest such value applies to all CPUs in any of the overlapping cpusets. In such cases, the value minus one (-1) is the lowest value, overridden by any other value, and the value zero (0) is the next lowest value.

Mask format¶

This format displays each 32-bit word in hexadecimal (using ASCII characters "0" - "9" and "a" - "f"); words are filled with leading zeros, if required. For masks longer than one word, a comma separator is used between words. Words are displayed in big-endian order, which has the most significant bit first. The hex digits within a word are also in big-endian order.

The number of 32-bit words displayed is the minimum number needed to display all bits of the bit mask, based on the size of the bit mask.

Examples of the Mask Format:

00000001                        # just bit 0 set
40000000,00000000,00000000      # just bit 94 set
00000001,00000000,00000000      # just bit 64 set
000000ff,00000000               # bits 32-39 set
00000000,000e3862               # 1,5,6,11-13,17-19 set

A mask with bits 0, 1, 2, 4, 8, 16, 32, and 64 set displays as:

00000001,00000001,00010117

The first "1" is for bit 64, the second for bit 32, the third for bit 16, the fourth for bit 8, the fifth for bit 4, and the "7" is for bits 2, 1, and 0.

List format¶

Examples of the List Format:

0-4,9           # bits 0, 1, 2, 3, 4, and 9 set
0-2,7,12-14     # bits 0, 1, 2, 7, 12, 13, and 14 set

Enabling memory_pressure¶

Using the echo command¶

echo 19 > cpuset.mems

failed because memory node 19 was not allowed (perhaps the current system does not have a memory node 19), then the echo command might not display any error. It is better to use the /bin/echo external command to change cpuset file settings, as this command will display write(2) errors, as in the example:

/bin/echo 19 > cpuset.mems
/bin/echo: write error: Invalid argument

Memory placement¶

If hot-plug functionality is used to remove all the CPUs that are currently assigned to a cpuset, then the kernel will automatically update the cpus_allowed of all processes attached to CPUs in that cpuset to allow all CPUs. When memory hot-plug functionality for removing memory nodes is available, a similar exception is expected to apply there as well. In general, the kernel prefers to violate cpuset placement, rather than starving a process that has had all its allowed CPUs or memory nodes taken offline. User code should reconfigure cpusets to refer only to online CPUs and memory nodes when using hot-plug to add or remove such resources.

A few kernel-critical, internal memory-allocation requests, marked GFP_ATOMIC, must be satisfied immediately. The kernel may drop some request or malfunction if one of these allocations fail. If such a request cannot be satisfied within the current process's cpuset, then we relax the cpuset, and look for memory anywhere we can find it. It's better to violate the cpuset than stress the kernel.

Allocations of memory requested by kernel drivers while processing an interrupt lack any relevant process context, and are not confined by cpusets.

Renaming cpusets¶

Creating and attaching to a cpuset.¶

1)

mkdir /dev/cpuset (if not already done)

2)

mount -t cpuset none /dev/cpuset (if not already done)

3)

Create the new cpuset using mkdir(1).

4)

Assign CPUs and memory nodes to the new cpuset.

5)

Attach the shell to the new cpuset.

For example, the following sequence of commands will set up a cpuset named "Charlie", containing just CPUs 2 and 3, and memory node 1, and then attach the current shell to that cpuset.

$ mkdir /dev/cpuset
$ mount -t cpuset cpuset /dev/cpuset
$ cd /dev/cpuset
$ mkdir Charlie
$ cd Charlie
$ /bin/echo 2-3 > cpuset.cpus
$ /bin/echo 1 > cpuset.mems
$ /bin/echo $$ > tasks
# The current shell is now running in cpuset Charlie
# The next line should display '/Charlie'
$ cat /proc/self/cpuset

Migrating a job to different memory nodes.¶

1)

Let's say we want to move the job in cpuset alpha (CPUs 4–7 and memory nodes 2–3) to a new cpuset beta (CPUs 16–19 and memory nodes 8–9).

2)

First create the new cpuset beta.

3)

Then allow CPUs 16–19 and memory nodes 8–9 in beta.

4)

Then enable memory_migration in beta.

5)

Then move each process from alpha to beta.

The following sequence of commands accomplishes this.

$ cd /dev/cpuset
$ mkdir beta
$ cd beta
$ /bin/echo 16-19 > cpuset.cpus
$ /bin/echo 8-9 > cpuset.mems
$ /bin/echo 1 > cpuset.memory_migrate
$ while read i; do /bin/echo $i; done < ../alpha/tasks > tasks

The above should move any processes in alpha to beta, and any memory held by these processes on memory nodes 2–3 to memory nodes 8–9, respectively.

Notice that the last step of the above sequence did not do:

$ cp ../alpha/tasks tasks

The while loop, rather than the seemingly easier use of the cp(1) command, was necessary because only one process PID at a time may be written to the tasks file.

The same effect (writing one PID at a time) as the while loop can be accomplished more efficiently, in fewer keystrokes and in syntax that works on any shell, but alas more obscurely, by using the -u (unbuffered) option of sed(1):

$ sed -un p < ../alpha/tasks > tasks