constant flow of new connection requests could prevent the postmaster from
completing a shutdown or crash restart. This is done by labeling child
processes that are "dead ends", that is, we know that they were launched only
to tell a client that it can't connect. These processes are managed
separately so that they don't confuse us into thinking that we can't advance
to the next stage of a shutdown or restart sequence, until the very end
where we must wait for them to drain out so we can delete the shmem segment.
Per discussion of a misbehavior reported by Keaton Adams.
Since this code was baroque already, and my first attempt at fixing the
problem made it entirely impenetrable, I took the opportunity to rewrite it
in a state-machine style. That eliminates some duplicated code sections and
hopefully makes everything a bit clearer.
as well as regular backends: if no regular backend launches before the autovac
launcher tries to start an autovac worker, the postmaster would get an Assert
fault due to calling PostmasterRandom before random_seed was initialized.
Cleanest solution seems to be to take the initialization of random_seed out
of ServerLoop and let PostmasterRandom do it for itself.
displayed in the postmaster log. This avoids Windows-specific problems with
localized time zone names that are in the wrong encoding, and generally seems
like a good idea to forestall other potential platform-dependent issues.
To preserve the existing behavior that all backends will log in the same time
zone, create a new GUC variable log_timezone that can only be changed on a
system-wide basis, and reference log-related calculations to that zone instead
of the TimeZone variable.
This fixes the issue reported by Hiroshi Saito that timestamps printed by
xlog.c startup could be improperly localized on Windows. We still need a
simpler patch for that problem in the back branches, however.
not bothering to initialize is_autovacuum for regular backends, meaning there
was a significant chance of the postmaster prematurely sending them SIGTERM
during database shutdown. Also, leaving the cancel key unset for an autovac
worker meant that any client could send it SIGINT, which doesn't sound
especially good either.
so that we will be able to create a cookie for all processes for CSVlogs.
It is set wherever MyProcPid is set. Take the opportunity to remove the now
unnecessary session-only restriction on the %s and %c escapes in log_line_prefix.
and fsync WAL at convenient intervals. For the moment it just tries to
offload this work from backends, but soon it will be responsible for
guaranteeing a maximum delay before asynchronously-committed transactions
will be flushed to disk.
This is a portion of Simon Riggs' async-commit patch, committed to CVS
separately because a background WAL writer seems like it might be a good idea
independently of the async-commit feature. I rebased walwriter.c on
bgwriter.c because it seemed like a more appropriate way of handling signals;
while the startup/shutdown logic in postmaster.c is more like autovac because
we want walwriter to quit before we start the shutdown checkpoint.
against a Unix server, and Windows-specific server-side authentication
using SSPI "negotiate" method (Kerberos or NTLM).
Only builds properly with MSVC for now.
we don't know at that point which relation OID to tell pgstat to forget.
The code was passing the relfilenode, which is incorrect, and could possibly
cause some other relation's stats to be zeroed out. While we could try to
clean this up, it seems much simpler and more reliable to let the next
invocation of pgstat_vacuum_tabstat() fix things; which indeed is how it
worked before I introduced the buggy code into 8.1.3 and later :-(.
Problem noticed by Itagaki Takahiro, fix is per subsequent discussion.
checkpoint. The comment claimed that we could do this anytime after
setting the checkpoint REDO point, but actually BufferSync is relying
on the assumption that buffers dumped by other backends will be fsync'd
too. So we really could not do it any sooner than we are doing it.
so that it responds to SIGQUIT reasonably promptly even on machines where
SA_RESTART signals restart a sleep from scratch. (This whole area could
stand some rethinking, but for now make it work like the other processes
do.) Also some marginal stylistic cleanups.
for it to die before telling the bgwriter to initiate shutdown checkpoint.
Since it's connected to shared memory, this seems more prudent than the
alternative of letting it quit asynchronously. Resolves my complaint
of yesterday about repeated shutdown checkpoints in CVS HEAD.
memory context pointing at a context not long lived enough.
Also, create a fake PortalContext where to store the vac_context, if only
to avoid having it be a top-level memory context.
continue with the schedule. Change current uses of SIGINT to abort a worker
into SIGTERM, which keeps the old behaviour of terminating the process.
Patch from ITAGAKI Takahiro, with some editorializing of my own.
over a fairly long period of time, rather than being spat out in a burst.
This happens only for background checkpoints carried out by the bgwriter;
other cases, such as a shutdown checkpoint, are still done at full speed.
Remove the "all buffers" scan in the bgwriter, and associated stats
infrastructure, since this seems no longer very useful when the checkpoint
itself is properly throttled.
Original patch by Itagaki Takahiro, reworked by Heikki Linnakangas,
and some minor API editorialization by me.
by having the postmaster signal it when certain failures occur. This requires
the postmaster setting a flag in shared memory, but should be as safe as the
pmsignal.c code is.
Also make sure the launcher honor's a postgresql.conf change turning it off
on SIGHUP.
reassembled in the syslogger before writing to the log file. This prevents
partial messages from being written, which mucks up log rotation, and
messages from different backends being interleaved, which causes garbled
logs. Backport as far as 8.0, where the syslogger was introduced.
Tom Lane and Andrew Dunstan
causes a division-by-zero error in the vacuum code. This can happen when there
are more workers than cost limit units.
Per report from Galy Lee in
<200705310914.l4V9E6JA094603@wwwmaster.postgresql.org>.
value for the vacuum code. Instead, make zero signify getting the value from a
higher level configuration facility, just like -1 in the original coding. We
still document that -1 is the value that disables the feature, to avoid
confusing the user unnecessarily.
Reported by Galy Lee in <200705310914.l4V9E6JA094603@wwwmaster.postgresql.org>;
per subsequent discussion.
buffers, rather than blowing out the whole shared-buffer arena. Aside from
avoiding cache spoliation, this fixes the problem that VACUUM formerly tended
to cause a WAL flush for every page it modified, because we had it hacked to
use only a single buffer. Those flushes will now occur only once per
ring-ful. The exact ring size, and the threshold for seqscans to switch into
the ring usage pattern, remain under debate; but the infrastructure seems
done. The key bit of infrastructure is a new optional BufferAccessStrategy
object that can be passed to ReadBuffer operations; this replaces the former
StrategyHintVacuum API.
This patch also changes the buffer usage-count methodology a bit: we now
advance usage_count when first pinning a buffer, rather than when last
unpinning it. To preserve the behavior that a buffer's lifetime starts to
decrease when it's released, the clock sweep code is modified to not decrement
usage_count of pinned buffers.
Work not done in this commit: teach GiST and GIN indexes to use the vacuum
BufferAccessStrategy for vacuum-driven fetches.
Original patch by Simon, reworked by Heikki and again by Tom.
or abort within a backend; rearrange InitPostgres processing to make it so.
Revealed by just-added Asserts along with ECPG regression tests (hm, I wonder
why the core regression tests didn't expose it?). This possibly is another
reason for missing stats updates ...
and aborted transactions have different effects; also teach it not to assume
that prepared transactions are always committed.
Along the way, simplify the pgstats API by tying counting directly to
Relations; I cannot detect any redeeming social value in having stats
pointers in HeapScanDesc and IndexScanDesc structures. And fix a few
corner cases in which counts might be missed because the relation's
pgstat_info pointer hadn't been set.
and TimestampDifference, to make coding clearer. I think this should also fix
the failure to start workers in platforms with low resolution timers, as
reported by Itagaki Takahiro.
messages to the stats collector. This avoids the problem that enabling
stats_row_level for autovacuum has a significant overhead for short
read-only transactions, as noted by Arjen van der Meijden. We can avoid
an extra gettimeofday call by piggybacking on the one done for WAL-logging
xact commit or abort (although that doesn't help read-only transactions,
since they don't WAL-log anything).
In my proposal for this, I noted that we could change the WAL log entries
for commit/abort to record full TimestampTz precision, instead of only
time_t as at present. That's not done in this patch, but will be committed
separately.
when a relation is opened multiple times in the same transaction. This is
particularly useful for system catalogs, which we may heap_open or index_open
many times in a transaction, and it doesn't really cost anything extra even
if the rel is touched but once. Motivated by study of an example from Greg
Stark, in which pgstat_initstats() accounted for an unreasonably large
fraction of the runtime.
processes to be running simultaneously. Also, now autovacuum processes do not
count towards the max_connections limit; they are counted separately from
regular processes, and are limited by the new GUC variable
autovacuum_max_workers.
The launcher now has intelligence to launch workers on each database every
autovacuum_naptime seconds, limited only on the max amount of worker slots
available.
Also, the global worker I/O utilization is limited by the vacuum cost-based
delay feature. Workers are "balanced" so that the total I/O consumption does
not exceed the established limit. This part of the patch was contributed by
ITAGAKI Takahiro.
Per discussion.
its table list and then rechecks pgstat before vacuuming each table to
verify that no one has vacuumed the table in the meantime.
In the current autovacuum world this only means that a worker will not
vacuum a table that a user has vacuumed manually after the worker started.
When support for multiple autovacuum workers is introduced, this will reduce
the probability of simultaneous workers on the same database doing redundant
work.
module and teach PREPARE and protocol-level prepared statements to use it.
In service of this, rearrange utility-statement processing so that parse
analysis does not assume table schemas can't change before execution for
utility statements (necessary because we don't attempt to re-acquire locks
for utility statements when reusing a stored plan). This requires some
refactoring of the ProcessUtility API, but it ends up cleaner anyway,
for instance we can get rid of the QueryContext global.
Still to do: fix up SPI and related code to use the plan cache; I'm tempted to
try to make SQL functions use it too. Also, there are at least some aspects
of system state that we want to ensure remain the same during a replan as in
the original processing; search_path certainly ought to behave that way for
instance, and perhaps there are others.
o read global SSL configuration file
o add GUC "ssl_ciphers" to control allowed ciphers
o add libpq environment variable PGSSLKEY to control SSL hardware keys
Victor B. Wagner
continuously, and requests vacuum runs of "autovacuum workers" to postmaster.
The workers do the actual vacuum work. This allows for future improvements,
like allowing multiple autovacuum jobs running in parallel.
For now, the code keeps the original behavior of having a single autovac
process at any time by sleeping until the previous worker has finished.
socket is still read-ready, the code was a tight loop, wasting lots of CPU.
We can't do anything to clear the failure, other than wait, but we should give
other processes more chance to finish and release FDs; so insert a small sleep.
Also, avoid bogus "close(-1)" in this case. Per report from Jim Nasby.
entries for the victim database go away sooner rather than later. We already
did the equivalent thing at the per-relation level, not sure why it's not
been done for whole databases. With this change, pgstat_vacuum_tabstat
should usually not find anything to do; though we still need it as a backstop
in case DROPDB or TABPURGE messages get lost under load.
already collected in the current transaction; this allows plpgsql functions to
watch for stats updates even though they are confined to a single transaction.
Use this instead of the previous kluge involving pg_stat_file() to wait for
the stats collector to update in the stats regression test. Internally,
decouple storage of stats snapshots from transaction boundaries; they'll
now stick around until someone calls pgstat_clear_snapshot --- which xact.c
still does at transaction end, to maintain the previous behavior. This makes
the logic a lot cleaner, at the price of a couple dozen cycles per transaction
exit.
"database system is ready to accept connections", which is issued by the
postmaster when it really is ready to accept connections. Per proposal from
Markus Schiltknecht and subsequent discussion.
input in the stats collector. Our select() emulation is apparently buggy
for UDP sockets :-(. This should resolve problems with stats collection
(and hence autovacuum) failing under more than minimal load. Diagnosis
and patch by Magnus Hagander.
Patch probably needs to be back-ported to 8.1 and 8.0, but first let's
see if it makes the buildfarm happy...
is deleted. A backend about to unlink a file now sends a "revoke fsync"
request to the bgwriter to make it clean out pending fsync requests. There
is still a race condition where the bgwriter may try to fsync after the unlink
has happened, but we can resolve that by rechecking the fsync request queue
to see if a revoke request arrived meanwhile. This eliminates the former
kluge of "just assuming" that an ENOENT failure is okay, and lets us handle
the fact that on Windows it might be EACCES too without introducing any
questionable assumptions. After an idea of mine improved by Magnus.
The HEAD patch doesn't apply cleanly to 8.2, but I'll see about a back-port
later. In the meantime this could do with some testing on Windows; I've been
able to force it through the code path via ENOENT, but that doesn't prove that
it actually fixes the Windows problem ...
accessing it, like DROP DATABASE. This allows the regression tests to pass
with autovacuum enabled, which open the gates for finally enabling autovacuum
by default.
(or other types of pg_class entry): the function pgstat_vacuum_tabstat,
invoked during VACUUM startup, had runtime proportional to the number of
stats table entries times the number of pg_class rows; in other words
O(N^2) if the stats collector's information is reasonably complete.
Replace list searching with a hash table to bring it back to O(N)
behavior. Per report from kim at myemma.com.
Back-patch as far as 8.1; 8.0 and before use different coding here.
an optarg). Add some comments noting that code in three different files has
to be kept in sync. Fix erroneous description of -S switch (it sets work_mem
not silent_mode), and do some light copy-editing elsewhere in postgres-ref.
identify long-running transactions. Since we already need to record
the transaction-start time (e.g. for now()), we don't need any
additional system calls to report this information.
Catversion bumped, initdb required.
should allow delete-pending files to actually go away, and thereby work
around the various complaints we've seen about 'permission denied'
errors in such cases. Should be reasonably harmless in any case...
StartupXLOG and ShutdownXLOG no longer need to be critical sections, because
in all contexts where they are invoked, elog(ERROR) would be translated to
elog(FATAL) anyway. (One change in bgwriter.c is needed to make this true:
set ExitOnAnyError before trying to exit. This is a good fix anyway since
the existing code would have gone into an infinite loop on elog(ERROR) during
shutdown.) That avoids a misleading report of PANIC during semi-orderly
failures. Modify the postmaster to include the startup process in the set of
processes that get SIGTERM when a fast shutdown is requested, and also fix it
to not try to restart the bgwriter if the bgwriter fails while trying to write
the shutdown checkpoint. Net result is that "pg_ctl stop -m fast" does
something reasonable for a system in warm standby mode, and so should Unix
system shutdown (ie, universal SIGTERM). Per gripe from Stephen Harris and
some corner-case testing of my own.
Windows), arrange for each postmaster child process to be its own process
group leader, and deliver signals SIGINT, SIGTERM, SIGQUIT to the whole
process group not only the direct child process. This provides saner behavior
for archive and recovery scripts; in particular, it's possible to shut down a
warm-standby recovery server using "pg_ctl stop -m immediate", since delivery
of SIGQUIT to the startup subprocess will result in killing the waiting
recovery_command. Also, this makes Query Cancel and statement_timeout apply
to scripts being run from backends via system(). (There is no support in the
core backend for that, but it's widely done using untrusted PLs.) Per gripe
from Stephen Harris and subsequent discussion.
promoted to FATAL) end in exit(1) not exit(0). Then change the postmaster to
allow exit(1) without a system-wide panic, but not for the startup subprocess
or the bgwriter. There were a couple of places that were using exit(1) to
deliberately force a system-wide panic; adjust these to be exit(2) instead.
This fixes the problem noted back in July that if the startup process exits
with elog(ERROR), the postmaster would think everything is hunky-dory and
proceed to start up. Alternative solutions such as trying to run the entire
startup process as a critical section seem less clean, primarily because of
the fact that a fair amount of startup code is shared by all postmaster
children in the EXEC_BACKEND case. We'd need an ugly special case somewhere
near the head of main.c to make it work if it's the child process's
responsibility to determine what happens; and what's the point when the
postmaster already treats different children differently?
in PITR scenarios. We now WAL-log the replacement of old XIDs with
FrozenTransactionId, so that such replacement is guaranteed to propagate to
PITR slave databases. Also, rather than relying on hint-bit updates to be
preserved, pg_clog is not truncated until all instances of an XID are known to
have been replaced by FrozenTransactionId. Add new GUC variables and
pg_autovacuum columns to allow management of the freezing policy, so that
users can trade off the size of pg_clog against the amount of freezing work
done. Revise the already-existing code that forces autovacuum of tables
approaching the wraparound point to make it more bulletproof; also, revise the
autovacuum logic so that anti-wraparound vacuuming is done per-table rather
than per-database. initdb forced because of changes in pg_class, pg_database,
and pg_autovacuum catalogs. Heikki Linnakangas, Simon Riggs, and Tom Lane.
pgstat_bestart() has been called; else any lock-block occurring
during InitPostgres() is disastrous. I believe this explains
recent wasp regression failure; at least it explains the crash I
got while trying to duplicate the problem. I also made
pgstat_report_activity() safe against the same scenario, just
in case. The report_waiting hazard was created by my patch of
19-Aug to include waiting status in pg_stat_activity.
that ps_status provides by appending 'waiting' to the PS display. This
completes the project of making it feasible to turn off process title
updates and instead rely on pg_stat_activity. Per my suggestion a few
weeks ago.
than N seconds apart. This allows a simple, if not very high performance,
means of guaranteeing that a PITR archive is no more than N seconds behind
real time. Also make pg_current_xlog_location return the WAL Write pointer,
add pg_current_xlog_insert_location to return the Insert pointer, and fix
pg_xlogfile_name_offset to return its results as a two-element record instead
of a smashed-together string, as per recent discussion.
Simon Riggs
such as debugging and performance measurement. This consists of two features:
a table of "rendezvous variables" that allows separately-loaded shared
libraries to communicate, and a new GUC setting "local_preload_libraries"
that allows libraries to be loaded into specific sessions without explicit
cooperation from the client application. To make local_preload_libraries
as flexible as possible, we do not restrict its use to superusers; instead,
it is restricted to load only libraries stored in $libdir/plugins/. The
existing LOAD command has also been modified to allow non-superusers to
LOAD libraries stored in this directory.
This patch also renames the existing GUC variable preload_libraries to
shared_preload_libraries (after a suggestion by Simon Riggs) and does some
code refactoring in dfmgr.c to improve clarity.
Korry Douglas, with a little help from Tom Lane.
loaded libraries: call functions _PG_init() and _PG_fini() if the library
defines such symbols. Hence we no longer need to specify an initialization
function in preload_libraries: we can assume that the library used the
_PG_init() convention, instead. This removes one source of pilot error
in use of preloaded libraries. Original patch by Ralf Engelschall,
preload_libraries changes by me.
(table or index) before trying to open its relcache entry. This fixes
race conditions in which someone else commits a change to the relation's
catalog entries while we are in process of doing relcache load. Problems
of that ilk have been reported sporadically for years, but it was not
really practical to fix until recently --- for instance, the recent
addition of WAL-log support for in-place updates helped.
Along the way, remove pg_am.amconcurrent: all AMs are now expected to support
concurrent update.
it's handled just about like timezone; in particular, don't try
to read anything during InitializeGUCOptions. Should solve current
startup failure on Windows, and avoid wasted cycles if a nondefault
setting is specified in postgresql.conf too. Possibly we need to
think about a more general solution for handling 'expensive to set'
GUC options.
pg_usleep at all. Instead call the replacement function in
port/win32/signal.c by that name. Avoids tricky macro-redefinition
logic and suppresses a compiler warning; furthermore it ensures that
no one can accidentally use the non-signal-aware version of pg_usleep
in a Windows backend.
Andrew Dunstan
Tom Lane
Magnus Hagander
Alvaro Herrera
Bruce Momjian
Peter Eisentraut
Neil Conway