Sears Russell
87ba38adb7
fix compilation when CONCURRENT_LRU is defined; make it the default (since it is only used by concurrentBufferManager)
2010-04-20 03:56:47 +00:00
Sears Russell
c7fa929e71
This fixes in Stasis issue 22.
...
Replacement policies now track the pin count of each page that they interact with. This behavior used to live in bufferHash, and had never been ported to concurrentBufferManager.
For various reasons, the pin counts have to be updated atomically with the LRU state. This would be very difficult to do inside of concurrentBufferManager, which sometimes needs to add and remove pages to replacement policies without holding any other latches.
2010-04-20 01:09:32 +00:00
Sears Russell
5138af6d1c
leave 2 pages in reserve for stale pin test (as opposed to one)
2010-04-19 23:09:30 +00:00
Sears Russell
887832582c
expose stasis_handle_t up through the buffer manager; add sequential I/O hinting; fix stupid busy-wait in prefetch
2010-04-12 20:54:10 +00:00
Sears Russell
7c5eb2a7e9
added dup method to io handles
2010-04-08 23:50:01 +00:00
Sears Russell
ff2f2e9e1e
add TregionPrefetch()
2010-03-31 22:41:33 +00:00
Sears Russell
358f037441
fix concurrency bug that cropped up when the concurrent buffer manager was used with the concurrent wrapper replacement policy. Concurrent wrapper hashes on pageid, but the buffer manager was changing the pageid out from under it
2010-03-31 22:39:07 +00:00
Sears Russell
d281282533
fix concurrentWrapper so that it implements the same api as the other replacement policies (it used to livelock when no pages were available. It now returns null).
2010-03-31 22:37:42 +00:00
Sears Russell
3c27ed566e
add prefetch primitives
2010-03-29 22:29:30 +00:00
Sears Russell
e17e948697
add missing return statement
2010-03-29 22:29:14 +00:00
Sears Russell
8a5bbd9a38
silence compiler warning
2010-03-29 22:28:32 +00:00
Sears Russell
6deb34f5b2
fix handling of loadPageForOperation during recovery; initialize rid.size in slottedLast
2010-03-22 19:56:55 +00:00
Sears Russell
85abeb9d77
Fix ArrayList regressions
2010-02-18 19:50:06 +00:00
Sears Russell
5ffd1110e9
Added concept of "transaction fingerprints"
...
The region allocator marks freed regions "Condemned" before actually freeing them. Condemned regions are associated
with a transaction; before they can be reused / coalesced the transaction that condemned them must complete.
The region allocator used "TisActiveTransaction(xid)" to determine whether a given transaction had completed. The
problem is that Stasis now reuses transaction id's, and condemned regions were never reclaimed in practice. A transaction
fingerprint is a unique identifier (currently transaction id, rec_lsn) that can be used to distinguish transactions that
did not execute concurrently.
I considered an alternative design based upon end-of-transaction actions. This would have worked as well, but would
have been a more invasive change.
2010-02-18 19:12:46 +00:00
Sears Russell
6454d55a42
fix bug in loadUninitializedPage(). It was passing INVALID_XID into the transaction table, which is not allowed
2010-02-18 19:04:39 +00:00
Sears Russell
4e100479d1
Add TarrayListDealloc() to arrayList. Remove one region allocation from each array list alloc. Add unit test for dealloc.
...
Also, removed some old compensations cruft.
2010-02-18 04:12:39 +00:00
Sears Russell
adfecfada3
Fix LSN handling in loadUninitializedPage()
2010-02-18 04:11:22 +00:00
Sears Russell
ba3179710b
Suppress shutdown warning when empty transactions are automatically rolled back.
2010-02-17 05:10:30 +00:00
Sears Russell
5d0500dc7a
Perl updates:
...
- Cleaned up initialization and shutdown.
- Add support for autocommitted hashes
- Wrote preliminary perldocs
- Updated sample cgi scripts.
- Added new (partial) wiki thing
- Added stasis.pl (a simple toplevel)
2010-02-17 05:05:23 +00:00
Sears Russell
98e0cb1fab
implement TregionSize(), add TregionForce()
2010-02-12 16:44:28 +00:00
Sears Russell
180fb12859
unit test for bug 22
2010-02-11 22:19:03 +00:00
Sears Russell
0afb9177ac
C++ compatibility fix
2010-02-11 22:18:49 +00:00
Sears Russell
c440f7ad62
fix typo in uncompiled code
2010-02-11 22:18:17 +00:00
Sears Russell
078d518bb6
cleaned up some lsn-free stuff
2010-02-11 21:21:31 +00:00
Sears Russell
757a877c84
added VALGRIND_MODE
2010-02-11 21:18:26 +00:00
Sears Russell
1ff9afb753
added new constant: INVALID_SIZE
2010-02-11 20:36:33 +00:00
Sears Russell
fac5d2ff37
fix default LSN handling for pages with headers
2010-02-11 20:35:38 +00:00
Sears Russell
a5abadb580
partial update for fastcgi, clean up Stasis.pm
2010-01-27 21:49:12 +00:00
Sears Russell
9bdaf23764
add some more thread-safe pblHt wrappers to fix lockManager
2010-01-23 00:27:01 +00:00
Sears Russell
031e21fa6c
fix race condition introduced by last commit
2010-01-22 23:52:57 +00:00
Sears Russell
3e2bbc2eb6
move lsn sets into reserve_entry. write_entry is now a no-op
2010-01-22 23:41:10 +00:00
Sears Russell
903728c589
shorten dirtyPageTable test.
2010-01-22 23:39:42 +00:00
Sears Russell
bb182ee0b0
fix new truncation check; it was using values for the old copy of the log to check the new copy.
2010-01-22 19:13:14 +00:00
Sears Russell
3849b5f131
silence compiler warnings
2010-01-22 19:09:58 +00:00
Sears Russell
49979ffa52
unit test for libredblack
2010-01-21 23:48:43 +00:00
Sears Russell
51dc17d1c3
fail earlier on mismatched truncation file. One run of check_multiplexer saw truncation leave a zero byte log
2010-01-21 23:48:04 +00:00
Sears Russell
bb92d86c85
fix check_lockManager. The lock manager now relies on a global mutex to protect its hashtables. Eventually, it should be ported over to the concurrentHash API (or something more general purpose...)
2010-01-21 22:43:38 +00:00
Sears Russell
434cec1b6a
fix for regression in r1335. nextUnstableLSN needs to point to the beginning of a valid log entry. Add sizeof(lsn_t) to its computation so that we point at the beginning of the next entry, not to its size
2010-01-21 22:30:02 +00:00
Sears Russell
17ac09c652
fix deadlock in check_multiplexer.
2010-01-21 21:53:18 +00:00
Sears Russell
e2498b62be
fix deadlock; reorderingHandle was acquiring a write latch while writing a log entry, leading to an inconsistent latch ordering
2010-01-21 21:05:25 +00:00
Sears Russell
dbf253275f
fix nested top action log race; because of a refactoring mistake, it was releasing a mutex without acquiring it...
2010-01-21 20:27:35 +00:00
Sears Russell
9dba8770f0
test TsoftCommit(); make sure it works when TforceCommits() is called, and when Tdeinit() is called
2010-01-21 18:57:58 +00:00
Sears Russell
5f705a4463
skip the log force when TforceCommits() is called and the entire log has already been forced. The old behavior generated a redundant CRC entry and force wrote it.
2010-01-21 18:42:50 +00:00
Sears Russell
9f0886a565
fix typo that disabled the new tests
2010-01-21 18:40:28 +00:00
Sears Russell
f2ee7b6573
fix bug exposed by new group commit behavior; nextUnstableLSN was returning the offset of the beginning of the CRC entry, not the end of it. This was causing an assertion to fail.
2010-01-21 18:28:34 +00:00
Sears Russell
250223086c
reordered includes for compatibility with gcc3
2010-01-21 00:03:17 +00:00
Sears Russell
f1f1d888e4
avoid use of -Wno-variadic-macros, which causes an error with gcc 3.
2010-01-20 23:51:54 +00:00
Sears Russell
b1f7d1947a
more log refactoring. safeWrites now holds a latch across reserve_entry, write_entry, and write_entry_done
2010-01-19 02:14:09 +00:00
Sears Russell
94b356f837
sample stasis web services for apache modperl and fastcgi. Both are multithreaded. However, the modperl one is racy. FastCGI is currently the recommended approach
2010-01-19 02:11:04 +00:00
Sears Russell
417eec45c4
harden Stasis.pm initialization against a wider range of crazy environments (ie: apache modperl and fastcgi). Add TsoftCommit and TforceCommits to the perl binding
2010-01-19 01:56:40 +00:00
