stasis-aries-wal/src/lladd/recovery2.c

288 lines
8.5 KiB
C
Raw Normal View History

2004-06-24 21:10:31 +00:00
/**
Replacement for recovery.c
A lot of refactoring has been done to simplify the contents of recovery.c
Hopefully, this file will be nice and clean. :)
*/
#include <pbl/pbl.h>
#include "linkedlist.h"
#include <lladd/logger/logger2.h>
#include <lladd/logger/logHandle.h>
#include <lladd/common.h>
#include <lladd/bufferManager.h>
#include <malloc.h>
#include <assert.h>
/** @todo This include is an artifact of our lack of infrastructure to support log iterator guards. */
#include <lladd/operations/prepare.h>
/** Maps from xid (int) to status:
NULL - Haven't seen XID
U - XID is scheduled for UNDO
C - XID committed, but not ended. (End records are written once
the XID is flushed to disk. They are only written during the
recovery process, and shutdown since LLADD follows a no force
policy.) @todo Implement end records.
@todo We don't ever read values out of this hash table... could it
be replaced entirely with transactionLSN?
*/
/*static pblHashTable_t * transactionStatus; */
static pblHashTable_t * transactionLSN;
static LinkedListPtr rollbackLSNs = NULL;
/**
Determines which transactions committed, and which need to be redone.
In the original version, this function also:
- Determined the point in the log at which to start the Redo pass.
- Calculated a list of all dirty pages.
It no longer does either of these things:
- A checkpointing algorithm could figure out where the redo pass
should begin. (It would then truncate the log at that point.) This
function could be called before analysis if efficiency is a concern.
- We were using the list of dirty pages as an optimization to prevent
the pages from being read later during recovery. Since this function
no longer reads the pages in, there's no longer any reason to build
the list of dirty pages.
*/
static void Analysis () {
LogEntry * e;
LogHandle lh = getLogHandle();
/** After recovery, we need to know what the highest XID in the
log was so that we don't accidentally reuse XID's. This keeps
track of that value. */
int highestXid = 0;
/** @todo loadCheckPoint() - Jump forward in the log to the last
checkpoint. (Maybe getLogHandle should do this automatically,
since the log will be truncated on checkpoint anyway.) */
while((e = nextInLog(&lh))) {
lsn_t * xactLSN = (lsn_t*)pblHtLookup(transactionLSN, &(e->xid), sizeof(int));
/* recordid rid = e->contents.update.rid; */
if(highestXid < e->xid) {
highestXid = e->xid;
}
/** Track LSN's in two data structures:
- map: xid -> max LSN
- sorted list of maxLSN's
*/
if(xactLSN == NULL) {
xactLSN = malloc(sizeof(lsn_t));
pblHtInsert(transactionLSN, &(e->xid), sizeof(int), xactLSN);
} else {
/* We've seen this xact before, and must have put a value in
rollbackLSNs for it. That value is now stale, so remove
it. */
DEBUG("Removing %ld\n", *xactLSN);
removeVal(&rollbackLSNs, *xactLSN);
}
/* Now, rollbackLSNs certainly does not contain an LSN for this
transaction, and *xactLSN points to a value in the hash, so
writing to it updates the hash. This doesn't update the
rollbackLSN data structure, so it doesn't hurt to update this
value for all log entries. */
*xactLSN = e->LSN;
switch(e->type) {
case XCOMMIT:
/* We've removed this XACT's last LSN from the list of LSN's to
be rolled back, so we're done. */
break;
case XEND:
/*
XEND means this transaction reached stable storage.
Therefore, we can skip redoing any of its operations. (The
timestamps on each page guarantee that the redo phase will
not overwrite this transaction's work with stale data.)
The redo phase checks for a transaction's presence in
transactionLSN before redoing its actions. Therefore, if
we remove this transaction from the transactionStatus hash,
it will not be redone.
*/
pblHtRemove(transactionLSN, &(e->xid), sizeof(int));
break;
case UPDATELOG:
case CLRLOG:
/*
If the last record we see for a transaction is an update or clr,
then the transaction must not have committed, so it must need
to be rolled back.
Add it to the appropriate list
*/
DEBUG("Adding %ld\n", e->LSN);
addSortedVal(&rollbackLSNs, e->LSN);
break;
case XABORT:
/* Don't want this XID in the list of rolled back lsn's since
this XACT will be rolled back during redo. */
break;
/* case XALLOC: */ /* @todo Don't use XALLOC anymore, dont need it here. */
/* assert (0); */
default:
assert (0);
}
free (e);
}
TsetXIDCount(highestXid);
}
static void Redo() {
LogHandle lh = getLogHandle();
LogEntry * e;
while((e = nextInLog(&lh))) {
/* int garbage; */
/* Check to see if this log entry is part of a transaction that needs to be redone. */
if(pblHtLookup(transactionLSN, &(e->xid), sizeof(int)) != NULL) {
/* Check to see if this log entry contains an action that needs to be redone. */
if(e->type == UPDATELOG ||
e->type == CLRLOG) {
/* redoOperation checks the page that contains e->rid, so we
don't need to check to see if the page is newer than this
log entry. */
redoUpdate(e);
}
}
}
}
static void Undo(int recovery) {
LogHandle lh;
void * prepare_guard_state;
/* printf("!"); fflush(NULL); */
prepare_guard_state = getPrepareGuardState();
while(rollbackLSNs != NULL) {
LogEntry * e;
lsn_t rollback = popMaxVal(&rollbackLSNs);
DEBUG("Undoing LSN %ld\n", (long int)rollback);
if(recovery) {
/** @todo shouldn't be hardcoded here! */
lh = getGuardedHandle(rollback, &prepareGuard, prepare_guard_state);
} else {
/** @todo probably want guards that are run during normal operation. */
lh = getLSNHandle(rollback);
}
e = readLSNEntry(rollback);
/* printf("#"); fflush(NULL); */
/* printf("e->prev_offset: %ld\n", e->prevLSN);
printf("prev_offset: %ld\n", lh.prev_offset); */
while((e = previousInTransaction(&lh))) {
lsn_t this_lsn, clr_lsn;
/* printf("."); fflush(NULL); */
switch(e->type) {
case UPDATELOG:
2004-06-24 21:10:31 +00:00
/* Sanity check. If this fails, we've already undone this
update, or something is wrong with the redo phase or normal operation. */
this_lsn= readLSN(e->contents.update.rid.page);
2004-06-24 21:10:31 +00:00
/* printf("1"); fflush(NULL); */
/* This check was incorrect. Since blobManager may lazily
update the pageManager, it is possible for this test to
fail. In that case, the undo is handled by blob manager
(it removes this page from it's dirty blob list), not
us. (Nope, it is now, again correct--fixed blobManager)*/
assert(e->LSN <= this_lsn);
2004-06-24 21:10:31 +00:00
/* printf("1a"); fflush(NULL); */
/* Need to log a clr here. */
clr_lsn = LogCLR(e);
/* writeLSN(clr_lsn, e->contents.update.rid.page); */
/* Undo update is a no-op if the page does not reflect this
update, but it will write the new clr_lsn. */
undoUpdate(e, clr_lsn);
2004-06-24 21:10:31 +00:00
/* printf("1b"); fflush(NULL); */
break;
case CLRLOG:
/* Don't need to do anything special to handle CLR's.
Iterator will correctly jump to clr's previous undo record. */
/* printf("2"); fflush(NULL); */
break;
/* case XALLOC: */
/* Don't use xalloc anymore. */
/*assert(0);*/
break;
case XABORT:
/* Since XABORT is a no-op, we can safely ignore it. (XABORT records may be passed in by undoTrans.)*/
break;
2004-06-24 21:10:31 +00:00
default:
printf ("Unknown log type to undo (TYPE=%d, XID= %d, LSN=%ld), skipping...\n", e->type, e->xid, e->LSN);
break;
}
}
/* printf("$"); fflush(NULL); */
}
free(prepare_guard_state);
}
void InitiateRecovery() {
transactionLSN = pblHtCreate();
DEBUG("Analysis started\n");
Analysis();
DEBUG("Redo started\n");
Redo();
DEBUG("Undo started\n");
Undo(1);
DEBUG("Recovery complete.\n");
/** @todo Should we manually empty the hash table? */
pblHtDelete(transactionLSN);
/** @todo CleanUp(); */
}
void undoTrans(TransactionLog transaction) {
if(rollbackLSNs) {
destroyList(rollbackLSNs);
}
rollbackLSNs = 0;
if(transaction.prevLSN > 0) {
/* printf("scheduling lsn %ld for undo.\n", transaction.prevLSN); */
addSortedVal(&rollbackLSNs, transaction.prevLSN);
} else {
/* Nothing to undo. (Happens for read-only xacts. */
}
Undo(0);
}