Movatterモバイル変換

18.4.1. Shared Memory and Semaphores

18.4.2. systemd RemoveIPC

18.4.3. Resource Limits

18.4.4. Linux Memory Overcommit

18.4.5. Linux Huge Pages

PostgreSQL requires the operating system to provide inter-process communication (IPC) features, specifically shared memory and semaphores. Unix-derived systems typically provide“System V”IPC,“POSIX”IPC, or both.Windows has its own implementation of these features and is not discussed here.

By default,PostgreSQL allocates a very small amount of System V shared memory, as well as a much larger amount of anonymousmmap shared memory. Alternatively, a single large System V shared memory region can be used (seeshared_memory_type). In addition a significant number of semaphores, which can be either System V or POSIX style, are created at server startup. Currently, POSIX semaphores are used on Linux and FreeBSD systems while other platforms use System V semaphores.

System VIPC features are typically constrained by system-wide allocation limits. WhenPostgreSQL exceeds one of these limits, the server will refuse to start and should leave an instructive error message describing the problem and what to do about it. (See alsoSection 18.3.1.) The relevant kernel parameters are named consistently across different systems;Table 18.1 gives an overview. The methods to set them, however, vary. Suggestions for some platforms are given below.

PostgreSQL requires a few bytes of System V shared memory (typically 48 bytes, on 64-bit platforms) for each copy of the server. On most modern operating systems, this amount can easily be allocated. However, if you are running many copies of the server or you explicitly configure the server to use large amounts of System V shared memory (seeshared_memory_type anddynamic_shared_memory_type), it may be necessary to increaseSHMALL, which is the total amount of System V shared memory system-wide. Note thatSHMALL is measured in pages rather than bytes on many systems.

Less likely to cause problems is the minimum size for shared memory segments (SHMMIN), which should be at most approximately 32 bytes forPostgreSQL (it is usually just 1). The maximum number of segments system-wide (SHMMNI) or per-process (SHMSEG) are unlikely to cause a problem unless your system has them set to zero.

When using System V semaphores,PostgreSQL uses one semaphore per allowed connection (max_connections), allowed autovacuum worker process (autovacuum_max_workers), allowed WAL sender process (max_wal_senders), and allowed background process (max_worker_processes), in sets of 19. Each such set will also contain a 20th semaphore which contains a“magic number”, to detect collision with semaphore sets used by other applications. The maximum number of semaphores in the system is set bySEMMNS, which consequently must be at least as high asmax_connections plusautovacuum_max_workers plusmax_wal_senders, plusmax_worker_processes, plus one extra for each 19 allowed connections plus workers (see the formula inTable 18.1). The parameterSEMMNI determines the limit on the number of semaphore sets that can exist on the system at one time. Hence this parameter must be at leastceil((max_connections + autovacuum_max_workers + max_wal_senders + max_worker_processes + 7) / 19). Lowering the number of allowed connections is a temporary workaround for failures, which are usually confusingly worded“No space left on device”, from the functionsemget.

In some cases it might also be necessary to increaseSEMMAP to be at least on the order ofSEMMNS. If the system has this parameter (many do not), it defines the size of the semaphore resource map, in which each contiguous block of available semaphores needs an entry. When a semaphore set is freed it is either added to an existing entry that is adjacent to the freed block or it is registered under a new map entry. If the map is full, the freed semaphores get lost (until reboot). Fragmentation of the semaphore space could over time lead to fewer available semaphores than there should be.

Various other settings related to“semaphore undo”, such asSEMMNU andSEMUME, do not affectPostgreSQL.

When using POSIX semaphores, the number of semaphores needed is the same as for System V, that is one semaphore per allowed connection (max_connections), allowed autovacuum worker process (autovacuum_max_workers), allowed WAL sender process (max_wal_senders), and allowed background process (max_worker_processes). On the platforms where this option is preferred, there is no specific kernel limit on the number of POSIX semaphores.

Ifsystemd is in use, some care must be taken that IPC resources (including shared memory) are not prematurely removed by the operating system. This is especially of concern when installing PostgreSQL from source. Users of distribution packages of PostgreSQL are less likely to be affected, as thepostgres user is then normally created as a system user.

The settingRemoveIPC inlogind.conf controls whether IPC objects are removed when a user fully logs out. System users are exempt. This setting defaults to on in stocksystemd, but some operating system distributions default it to off.

A typical observed effect when this setting is on is that shared memory objects used for parallel query execution are removed at apparently random times, leading to errors and warnings while attempting to open and remove them, like

WARNING:  could not remove shared memory segment "/PostgreSQL.1450751626": No such file or directory

Different types of IPC objects (shared memory vs. semaphores, System V vs. POSIX) are treated slightly differently bysystemd, so one might observe that some IPC resources are not removed in the same way as others. But it is not advisable to rely on these subtle differences.

A“user logging out” might happen as part of a maintenance job or manually when an administrator logs in as thepostgres user or something similar, so it is hard to prevent in general.

What is a“system user” is determined atsystemd compile time from theSYS_UID_MAX setting in/etc/login.defs.

Packaging and deployment scripts should be careful to create thepostgres user as a system user by usinguseradd -r,adduser --system, or equivalent.

Alternatively, if the user account was created incorrectly or cannot be changed, it is recommended to set

RemoveIPC=no

in/etc/systemd/logind.conf or another appropriate configuration file.

Unix-like operating systems enforce various kinds of resource limits that might interfere with the operation of yourPostgreSQL server. Of particular importance are limits on the number of processes per user, the number of open files per process, and the amount of memory available to each process. Each of these have a“hard” and a“soft” limit. The soft limit is what actually counts but it can be changed by the user up to the hard limit. The hard limit can only be changed by the root user. The system callsetrlimit is responsible for setting these parameters. The shell's built-in commandulimit (Bourne shells) orlimit (csh) is used to control the resource limits from the command line. On BSD-derived systems the file/etc/login.conf controls the various resource limits set during login. See the operating system documentation for details. The relevant parameters aremaxproc,openfiles, anddatasize. For example:

default:\...        :datasize-cur=256M:\        :maxproc-cur=256:\        :openfiles-cur=256:\...

(-cur is the soft limit. Append-max to set the hard limit.)

Kernels can also have system-wide limits on some resources.

ThePostgreSQL server uses one process per connection so you should provide for at least as many processes as allowed connections, in addition to what you need for the rest of your system. This is usually not a problem but if you run several servers on one machine things might get tight.

The factory default limit on open files is often set to“socially friendly” values that allow many users to coexist on a machine without using an inappropriate fraction of the system resources. If you run many servers on a machine this is perhaps what you want, but on dedicated servers you might want to raise this limit.

On the other side of the coin, some systems allow individual processes to open large numbers of files; if more than a few processes do so then the system-wide limit can easily be exceeded. If you find this happening, and you do not want to alter the system-wide limit, you can setPostgreSQL'smax_files_per_process configuration parameter to limit the consumption of open files.

Another kernel limit that may be of concern when supporting large numbers of client connections is the maximum socket connection queue length. If more than that many connection requests arrive within a very short period, some may get rejected before thePostgreSQL server can service the requests, with those clients receiving unhelpful connection failure errors such as“Resource temporarily unavailable” or“Connection refused”. The default queue length limit is 128 on many platforms. To raise it, adjust the appropriate kernel parameter viasysctl, then restart thePostgreSQL server. The parameter is variously namednet.core.somaxconn on Linux,kern.ipc.soacceptqueue on newer FreeBSD, andkern.ipc.somaxconn on macOS and other BSD variants.

The default virtual memory behavior on Linux is not optimal forPostgreSQL. Because of the way that the kernel implements memory overcommit, the kernel might terminate thePostgreSQL postmaster (the supervisor server process) if the memory demands of eitherPostgreSQL or another process cause the system to run out of virtual memory.

If this happens, you will see a kernel message that looks like this (consult your system documentation and configuration on where to look for such a message):

Out of Memory: Killed process 12345 (postgres).

This indicates that thepostgres process has been terminated due to memory pressure. Although existing database connections will continue to function normally, no new connections will be accepted. To recover,PostgreSQL will need to be restarted.

One way to avoid this problem is to runPostgreSQL on a machine where you can be sure that other processes will not run the machine out of memory. If memory is tight, increasing the swap space of the operating system can help avoid the problem, because the out-of-memory (OOM) killer is invoked only when physical memory and swap space are exhausted.

IfPostgreSQL itself is the cause of the system running out of memory, you can avoid the problem by changing your configuration. In some cases, it may help to lower memory-related configuration parameters, particularlyshared_buffers,work_mem, andhash_mem_multiplier. In other cases, the problem may be caused by allowing too many connections to the database server itself. In many cases, it may be better to reducemax_connections and instead make use of external connection-pooling software.

It is possible to modify the kernel's behavior so that it will not“overcommit” memory. Although this setting will not prevent theOOM killer from being invoked altogether, it will lower the chances significantly and will therefore lead to more robust system behavior. This is done by selecting strict overcommit mode viasysctl:

sysctl -w vm.overcommit_memory=2

or placing an equivalent entry in/etc/sysctl.conf. You might also wish to modify the related settingvm.overcommit_ratio. For details see the kernel documentation filehttps://www.kernel.org/doc/Documentation/vm/overcommit-accounting.

Another approach, which can be used with or without alteringvm.overcommit_memory, is to set the process-specificOOM score adjustment value for the postmaster process to-1000, thereby guaranteeing it will not be targeted by the OOM killer. The simplest way to do this is to execute

echo -1000 > /proc/self/oom_score_adj

in thePostgreSQL startup script just before invokingpostgres. Note that this action must be done as root, or it will have no effect; so a root-owned startup script is the easiest place to do it. If you do this, you should also set these environment variables in the startup script before invokingpostgres:

export PG_OOM_ADJUST_FILE=/proc/self/oom_score_adjexport PG_OOM_ADJUST_VALUE=0

These settings will cause postmaster child processes to run with the normal OOM score adjustment of zero, so that the OOM killer can still target them at need. You could use some other value forPG_OOM_ADJUST_VALUE if you want the child processes to run with some other OOM score adjustment. (PG_OOM_ADJUST_VALUE can also be omitted, in which case it defaults to zero.) If you do not setPG_OOM_ADJUST_FILE, the child processes will run with the same OOM score adjustment as the postmaster, which is unwise since the whole point is to ensure that the postmaster has a preferential setting.

Using huge pages reduces overhead when using large contiguous chunks of memory, asPostgreSQL does, particularly when using large values ofshared_buffers. To use this feature inPostgreSQL you need a kernel withCONFIG_HUGETLBFS=y andCONFIG_HUGETLB_PAGE=y. You will also have to configure the operating system to provide enough huge pages of the desired size. The runtime-computed parametershared_memory_size_in_huge_pages reports the number of huge pages required. This parameter can be viewed before starting the server with apostgres command like:

$postgres -D $PGDATA -C shared_memory_size_in_huge_pages3170$grep ^Hugepagesize /proc/meminfoHugepagesize:       2048 kB$ls /sys/kernel/mm/hugepageshugepages-1048576kB  hugepages-2048kB

In this example the default is 2MB, but you can also explicitly request either 2MB or 1GB withhuge_page_size to adapt the number of pages calculated byshared_memory_size_in_huge_pages. While we need at least3170 huge pages in this example, a larger setting would be appropriate if other programs on the machine also need huge pages. We can set this with:

#sysctl -w vm.nr_hugepages=3170

Don't forget to add this setting to/etc/sysctl.conf so that it is reapplied after reboots. For non-default huge page sizes, we can instead use:

#echo 3170 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages

It is also possible to provide these settings at boot time using kernel parameters such ashugepagesz=2M hugepages=3170.

Sometimes the kernel is not able to allocate the desired number of huge pages immediately due to fragmentation, so it might be necessary to repeat the command or to reboot. (Immediately after a reboot, most of the machine's memory should be available to convert into huge pages.) To verify the huge page allocation situation for a given size, use:

$cat /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages

It may also be necessary to give the database server's operating system user permission to use huge pages by settingvm.hugetlb_shm_group viasysctl, and/or give permission to lock memory withulimit -l.

The default behavior for huge pages inPostgreSQL is to use them when possible, with the system's default huge page size, and to fall back to normal pages on failure. To enforce the use of huge pages, you can sethuge_pages toon inpostgresql.conf. Note that with this settingPostgreSQL will fail to start if not enough huge pages are available.

For a detailed description of theLinux huge pages feature have a look athttps://www.kernel.org/doc/Documentation/vm/hugetlbpage.txt.