The flags to the driver use new connections for new LBs since
fda4d46fc4. This makes it consistent with what we do for
MySQL already.
This should fix warnings about TransactionProfiler in objectcache,
as well as warnings about "Pending writes" in WANObjectCache.
Bug: T167946
Bug: T154424
Change-Id: I0b0d9a7210b6a3270d32df778fcc4b9918d3dcd1
The WMF parser cache tables have some 30000 rows with an identical, old
expiry time. So each time deleteObjectsExpiringBefore() is run, the
first batch on each table is correct, but the keys for the second batch
are selected with exptime > X, where X is shared exptime of the many old
rows.
So, use exptime >= X instead.
Change-Id: I6853f64f88f65d4529be2a42c9ed70cfa62cf653
It's unreasonable to expect newbies to know that "bug 12345" means "Task T14345"
except where it doesn't, so let's just standardise on the real numbers.
Change-Id: I6f59febaf8fc96e80f8cfc11f4356283f461142a
While prepping the code base for using etsy/phan i found a few problems.
A couple are real problems, while others are just places etsy couldn't
figure out the types.
* AuthPlugin referenced UserLoginTemplate which doesn't exist. Tracing
back usages this accepts classes that extend from BaseTemplate.
* DatabaseSqlite had an invalid @var annotation.
* LoadBalancer had some missing or under-defined annotations. Make them
more specific where possible.
* SqlBagOStuff didn't have the appropriate use statement for
WaitConditionLoop and probably doesn't work. Also updated annotations
that point to non-existent DatabaseBase to use IDatabase.
Change-Id: Iff2270b418ad2f8f97cfdb6df646c435d3283536
* Use OutputPage::output() as the method to mask latency, since it
takes a good while to run. By the time it runs, cache replication
should have caught up, so the reap call will likely not block.
* For redirects emitted after changes in POST, instead of masking
with OutputPage, add a parameter to the redirect and block on
the positions appearing. This uses the redirection RTT to mask
the replication latency.
Change-Id: Ib23690c302e8033610fef9a0ef451dafe8a5803e
This is more consistent with LoadBalancer, modern, and inclusive
of master/master mysql, NDB cluster, and MariaDB galera cluster.
The old constant is an alias now.
Change-Id: I0b37299ecb439cc446ffbe8c341365d1eef45849
* Keep track of the custom LoadBalancer when it makes one.
* Use the custom LoadBalancer to wait for slaves if one
was used, rather than the main singleton.
* Only wait on the slaves in the LoadBalancer if the
main DBs are being used.
Change-Id: I11de814306c44f27e0c33b08b5921c0fd4cdc24f
* Refactor local DB usage check into usesMainDB() method.
* Avoid using the db member of DBError instances.
Change-Id: I7350f5a471c551492094bfaf545ebc222eb6f7dd
Rely on the mass-rollback logic in MWExceptionHandler instead.
This results in a better chance of atomicity.
Change-Id: I2eb5661d4acc105a1323d69c5463268c234bd745
* Add the ability to expose key BagOStuff attributes like whether the
emulation SQL cache is being used. Several callers use hacks to detect
this and can be updated later.
* Fallback to a safe empty WAN cache in the CACHE_DB case. This fixes
a regression from f4bf52e843.
* Also add this protection to the server cache, which could break
similarly before too.
* Also fix CACHE_ANYTHING by avoid the recursion risk from
fc1d4d7960 by just checking the disabled service map.
Bug: T123829
Bug: T141804
Change-Id: I17ee26138f69e01ec1aaddb55ab27caa4d542193
Mostly places which immediately had a:
if ( !is_callable( $callback ) ) {
throw new Exception(...);
}
check at the beginning of the function.
Change-Id: Ia78663b2231629010816bd1cda8814b996968d1d
* This blocks on writing to all replicas
and returns false if any failed.
* This is useful if ChronologyProtector is to work across
domains by having the writes go everywhere so that later
reads will see them (and be local at the same time).
* Redundant doc comments were also removed.
Change-Id: I9ed098d563c64dba605e7809bc96731da3b3e79d
* Callers of get() no longer have to contend with
the annoying $casToken parameter, which is there
but totally unusable to non-BagOStuff code.
* The default get() now delegates to doGet(),
which callers must implement instead. They can
ignore the overhead of generating $casToken if
they do not implement cas(), which applies to
callers that use the stock merge(). If cas() is
used for merge(), then getWithToken() must be
implemented.
* Also add BagOStuff::READ_LATEST to mergeViaCas()
for sanity, as that missing before.
Likewise with mergeViaLock().
Change-Id: I4efce6a9ab4b1eadd2f161dff641004a7239c516
* Simplify the debug log call and use queries group
* Remove $wgDebugDumpSqlLength, as profiler output
already has shortened query strings (one can use
profiling without DBO_DEBUG)
* Removed $wgDebugDBTransactions as BEGIN/COMMIT already show
* Removed PostgresTransactionState as it was only used for
$wgDebugDBTransactions handling
* This cuts down on lots of global variable usage
Change-Id: I185adb1694441d074dea965960429b4910727620
* This also makes sure garbageCollect() is called on writes from
time to time, since set() is used much more often
Change-Id: I5b6cc686bd6204fc3d6ff8b003be8b8d6a493e24
* To make this work, a slaveOnly flag was added to SqlBagOStuff
and it no longer prunes expired items on get(), it just treats
them as missing. The use of garbageCollect() for the writeFactory
cache and the use of replace() on updates already makes this
a non-issue.
* This also makes the default $wgMainStash non-bogus.
Bug: T88493
Change-Id: I9d1f31305e08430de29a3cd521cdb10a82dffd10
* A "gone away" error can trigger ping(), which can fail,
triggering a full open(). If this fails (as it often will),
mConn is false and the DatabaseBase object is not usable.
Bug: T103435
Change-Id: Id2aaed87d2642e73c7811076da908180ee5ab5a4
- Removed space after casts
- Removed spaces in array index
- Added spaces around string concat
- Added space after words: switch, foreach
- else if -> elseif
- Removed parentheses around require_once, because it is not a function
- Added newline at end of file
- Removed double spaces
- Added spaces around operations
- Removed repeated newlines
Bug: T102609
Change-Id: Ib860222b24f8ad8e9062cd4dc42ec88dc63fb49e
wfSuppressWarnings() and wfRestoreWarnings() were split out into a
separate library. All usages in core were replaced with the new
functions, and the wf* global functions are marked as deprecated.
Additionally, some uses of @ were replaced due to composer's autoloader
being loaded even earlier.
Ie1234f8c12693408de9b94bf6f84480a90bd4f8e adds the library to
mediawiki/vendor.
Bug: T100923
Change-Id: I5c35079a0a656180852be0ae6b1262d40f6534c4
* Set a custom profiler with no expectations to avoid tripping
the "0 write" expection. This avoids useless log entries.
Change-Id: Iac849a729eb36b1a8affb0dbc8b8c195fab4b03a
* merge() using the locking version by default. The callers that implement cas()
override merge() to use the cas-based version. Those that do not no longer
need to define the unused dummy method.
* Removed some redundant doc blocks.
Change-Id: I49f7dd5432efa8d76f4758c273f0859376ddafb7
ObjectCache::newFromParams() will handle a 'loggroup' parameter
specially, getting an instance of \Psr\Log\LoggerInterface to
pass to the BagsOStuff.
BagOStuff implementations can use $this->logger which will be
an implementation of \Psr\Log\LoggerInterface. As this is set in
BagOStuff::__construct(), all subclasses must now call the
parent constructor.
The goal of this is to make the logging in BagOStuff non-MediaWiki
specific, in the hopes of separating it out into a separate library
in the future.
Change-Id: I8a8e278e6f028814499d8457d6d5341d03eabc7a
