/*
* pg_prewarm(regclass, mode text, fork text,
* first_block int8, last_block int8)
*
* The first argument is the relation to be prewarmed; the second controls
* how prewarming is done; legal options are 'prefetch', 'read', and 'buffer'.
* The third is the name of the relation fork to be prewarmed. The fourth
* and fifth arguments specify the first and last block to be prewarmed.
* If the fourth argument is NULL, it will be taken as 0; if the fifth argument
* is NULL, it will be taken as the number of blocks in the relation. The
* return value is the number of blocks successfully prewarmed.
*/
Datum
pg_prewarm(PG_FUNCTION_ARGS)
{
Oid relOid;
text *forkName;
text *type;
int64 first_block;
int64 last_block;
int64 nblocks;
int64 blocks_done = 0;
int64 block;
Relation rel;
ForkNumber forkNumber;
char *forkString;
char *ttype;
PrewarmType ptype;
AclResult aclresult;
/* Basic sanity checking. */
if (PG_ARGISNULL(0))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("relation cannot be null")));
relOid = PG_GETARG_OID(0);
if (PG_ARGISNULL(1))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
(errmsg("prewarm type cannot be null"))));
type = PG_GETARG_TEXT_P(1);
ttype = text_to_cstring(type);
if (strcmp(ttype, "prefetch") == 0)
ptype = PREWARM_PREFETCH;
else if (strcmp(ttype, "read") == 0)
ptype = PREWARM_READ;
else if (strcmp(ttype, "buffer") == 0)
ptype = PREWARM_BUFFER;
else
{
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid prewarm type"),
errhint("Valid prewarm types are \"prefetch\", \"read\", and \"buffer\".")));
PG_RETURN_INT64(0); /* Placate compiler. */
}
if (PG_ARGISNULL(2))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
(errmsg("relation fork cannot be null"))));
forkName = PG_GETARG_TEXT_P(2);
forkString = text_to_cstring(forkName);
forkNumber = forkname_to_number(forkString);
/* Open relation and check privileges. */
rel = relation_open(relOid, AccessShareLock);
aclresult = pg_class_aclcheck(relOid, GetUserId(), ACL_SELECT);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_CLASS, get_rel_name(relOid));
/* Check that the fork exists. */
RelationOpenSmgr(rel);
if (!smgrexists(rel->rd_smgr, forkNumber))
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("fork \"%s\" does not exist for this relation",
forkString)));
/* Validate block numbers, or handle nulls. */
nblocks = RelationGetNumberOfBlocksInFork(rel, forkNumber);
if (PG_ARGISNULL(3))
first_block = 0;
else
{
first_block = PG_GETARG_INT64(3);
if (first_block < 0 || first_block >= nblocks)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("starting block number must be between 0 and " INT64_FORMAT,
nblocks - 1)));
}
if (PG_ARGISNULL(4))
last_block = nblocks - 1;
else
{
last_block = PG_GETARG_INT64(4);
if (last_block < 0 || last_block >= nblocks)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("ending block number must be between 0 and " INT64_FORMAT,
nblocks - 1)));
}
/* Now we're ready to do the real work. */
if (ptype == PREWARM_PREFETCH)
{
#ifdef USE_PREFETCH
/*
* In prefetch mode, we just hint the OS to read the blocks, but we
* don't know whether it really does it, and we don't wait for it to
* finish.
*
* It would probably be better to pass our prefetch requests in chunks
* of a megabyte or maybe even a whole segment at a time, but there's
* no practical way to do that at present without a gross modularity
* violation, so we just do this.
*/
for (block = first_block; block <= last_block; ++block)
{
CHECK_FOR_INTERRUPTS();
PrefetchBuffer(rel, forkNumber, block);
++blocks_done;
}
#else
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("prefetch is not supported by this build")));
#endif
}
else if (ptype == PREWARM_READ)
{
/*
* In read mode, we actually read the blocks, but not into shared
* buffers. This is more portable than prefetch mode (it works
* everywhere) and is synchronous.
*/
for (block = first_block; block <= last_block; ++block)
{
CHECK_FOR_INTERRUPTS();
smgrread(rel->rd_smgr, forkNumber, block, blockbuffer.data);
++blocks_done;
}
}
else if (ptype == PREWARM_BUFFER)
{
/*
* In buffer mode, we actually pull the data into shared_buffers.
*/
for (block = first_block; block <= last_block; ++block)
{
Buffer buf;
CHECK_FOR_INTERRUPTS();
buf = ReadBufferExtended(rel, forkNumber, block, RBM_NORMAL, NULL);
ReleaseBuffer(buf);
++blocks_done;
}
}
/* Close relation, release lock. */
relation_close(rel, AccessShareLock);
PG_RETURN_INT64(blocks_done);
}
Datum
pgrowlocks(PG_FUNCTION_ARGS)
{
FuncCallContext *funcctx;
HeapScanDesc scan;
HeapTuple tuple;
TupleDesc tupdesc;
AttInMetadata *attinmeta;
Datum result;
MyData *mydata;
Relation rel;
if (SRF_IS_FIRSTCALL())
{
text *relname;
RangeVar *relrv;
MemoryContext oldcontext;
AclResult aclresult;
funcctx = SRF_FIRSTCALL_INIT();
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
attinmeta = TupleDescGetAttInMetadata(tupdesc);
funcctx->attinmeta = attinmeta;
relname = PG_GETARG_TEXT_P(0);
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
rel = heap_openrv(relrv, AccessShareLock);
/* check permissions: must have SELECT on table */
aclresult = pg_class_aclcheck(RelationGetRelid(rel), GetUserId(),
ACL_SELECT);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_CLASS,
RelationGetRelationName(rel));
scan = heap_beginscan(rel, SnapshotNow, 0, NULL);
mydata = palloc(sizeof(*mydata));
mydata->rel = rel;
mydata->scan = scan;
mydata->ncolumns = tupdesc->natts;
funcctx->user_fctx = mydata;
MemoryContextSwitchTo(oldcontext);
}
funcctx = SRF_PERCALL_SETUP();
attinmeta = funcctx->attinmeta;
mydata = (MyData *) funcctx->user_fctx;
scan = mydata->scan;
/* scan the relation */
while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
{
/* must hold a buffer lock to call HeapTupleSatisfiesUpdate */
LockBuffer(scan->rs_cbuf, BUFFER_LOCK_SHARE);
if (HeapTupleSatisfiesUpdate(rel, tuple->t_data,
GetCurrentCommandId(/*false*/),
scan->rs_cbuf) == HeapTupleBeingUpdated)
{
char **values;
int i;
values = (char **) palloc(mydata->ncolumns * sizeof(char *));
i = 0;
values[i++] = (char *) DirectFunctionCall1(tidout, PointerGetDatum(&tuple->t_self));
if (tuple->t_data->t_infomask & HEAP_XMAX_SHARED_LOCK)
values[i++] = pstrdup("Shared");
else
values[i++] = pstrdup("Exclusive");
values[i] = palloc(NCHARS * sizeof(char));
snprintf(values[i++], NCHARS, "%d", HeapTupleHeaderGetXmax(tuple->t_data));
if (tuple->t_data->t_infomask & HEAP_XMAX_IS_MULTI)
{
TransactionId *xids;
int nxids;
int j;
int isValidXid = 0; /* any valid xid ever exists? */
values[i++] = pstrdup("true");
nxids = GetMultiXactIdMembers(HeapTupleHeaderGetXmax(tuple->t_data), &xids);
if (nxids == -1)
{
elog(ERROR, "GetMultiXactIdMembers returns error");
}
values[i] = palloc(NCHARS * nxids);
values[i + 1] = palloc(NCHARS * nxids);
strcpy(values[i], "{");
strcpy(values[i + 1], "{");
for (j = 0; j < nxids; j++)
{
char buf[NCHARS];
if (TransactionIdIsInProgress(xids[j]))
{
if (isValidXid)
{
strcat(values[i], ",");
strcat(values[i + 1], ",");
}
snprintf(buf, NCHARS, "%d", xids[j]);
strcat(values[i], buf);
snprintf(buf, NCHARS, "%d", BackendXidGetPid(xids[j]));
strcat(values[i + 1], buf);
isValidXid = 1;
}
}
strcat(values[i], "}");
strcat(values[i + 1], "}");
i++;
}
else
{
values[i++] = pstrdup("false");
values[i] = palloc(NCHARS * sizeof(char));
snprintf(values[i++], NCHARS, "{%d}", HeapTupleHeaderGetXmax(tuple->t_data));
values[i] = palloc(NCHARS * sizeof(char));
snprintf(values[i++], NCHARS, "{%d}", BackendXidGetPid(HeapTupleHeaderGetXmax(tuple->t_data)));
}
LockBuffer(scan->rs_cbuf, BUFFER_LOCK_UNLOCK);
/* build a tuple */
tuple = BuildTupleFromCStrings(attinmeta, values);
/* make the tuple into a datum */
result = HeapTupleGetDatum(tuple);
/* Clean up */
for (i = 0; i < mydata->ncolumns; i++)
pfree(values[i]);
pfree(values);
SRF_RETURN_NEXT(funcctx, result);
}
else
{
LockBuffer(scan->rs_cbuf, BUFFER_LOCK_UNLOCK);
}
}
heap_endscan(scan);
heap_close(mydata->rel, AccessShareLock);
SRF_RETURN_DONE(funcctx);
}
/*
* BuildIndexValueDescription
*
* Construct a string describing the contents of an index entry, in the
* form "(key_name, ...)=(key_value, ...)". This is currently used
* for building unique-constraint and exclusion-constraint error messages.
*
* Note that if the user does not have permissions to view all of the
* columns involved then a NULL is returned. Returning a partial key seems
* unlikely to be useful and we have no way to know which of the columns the
* user provided (unlike in ExecBuildSlotValueDescription).
*
* The passed-in values/nulls arrays are the "raw" input to the index AM,
* e.g. results of FormIndexDatum --- this is not necessarily what is stored
* in the index, but it's what the user perceives to be stored.
*/
char *
BuildIndexValueDescription(Relation indexRelation,
Datum *values, bool *isnull)
{
StringInfoData buf;
Form_pg_index idxrec;
HeapTuple ht_idx;
int natts = indexRelation->rd_rel->relnatts;
int i;
int keyno;
Oid indexrelid = RelationGetRelid(indexRelation);
Oid indrelid;
AclResult aclresult;
/*
* Check permissions- if the user does not have access to view all of the
* key columns then return NULL to avoid leaking data.
*
* First check if RLS is enabled for the relation. If so, return NULL to
* avoid leaking data.
*
* Next we need to check table-level SELECT access and then, if there is
* no access there, check column-level permissions.
*/
/*
* Fetch the pg_index tuple by the Oid of the index
*/
ht_idx = SearchSysCache1(INDEXRELID, ObjectIdGetDatum(indexrelid));
if (!HeapTupleIsValid(ht_idx))
elog(ERROR, "cache lookup failed for index %u", indexrelid);
idxrec = (Form_pg_index) GETSTRUCT(ht_idx);
indrelid = idxrec->indrelid;
Assert(indexrelid == idxrec->indexrelid);
/* RLS check- if RLS is enabled then we don't return anything. */
if (check_enable_rls(indrelid, InvalidOid, true) == RLS_ENABLED)
{
ReleaseSysCache(ht_idx);
return NULL;
}
/* Table-level SELECT is enough, if the user has it */
aclresult = pg_class_aclcheck(indrelid, GetUserId(), ACL_SELECT);
if (aclresult != ACLCHECK_OK)
{
/*
* No table-level access, so step through the columns in the index and
* make sure the user has SELECT rights on all of them.
*/
for (keyno = 0; keyno < idxrec->indnatts; keyno++)
{
AttrNumber attnum = idxrec->indkey.values[keyno];
/*
* Note that if attnum == InvalidAttrNumber, then this is an index
* based on an expression and we return no detail rather than try
* to figure out what column(s) the expression includes and if the
* user has SELECT rights on them.
*/
if (attnum == InvalidAttrNumber ||
pg_attribute_aclcheck(indrelid, attnum, GetUserId(),
ACL_SELECT) != ACLCHECK_OK)
{
/* No access, so clean up and return */
ReleaseSysCache(ht_idx);
return NULL;
}
}
}
ReleaseSysCache(ht_idx);
initStringInfo(&buf);
appendStringInfo(&buf, "(%s)=(",
pg_get_indexdef_columns(indexrelid, true));
for (i = 0; i < natts; i++)
{
char *val;
if (isnull[i])
val = "null";
else
{
Oid foutoid;
bool typisvarlena;
/*
* The provided data is not necessarily of the type stored in the
* index; rather it is of the index opclass's input type. So look
* at rd_opcintype not the index tupdesc.
*
* Note: this is a bit shaky for opclasses that have pseudotype
* input types such as ANYARRAY or RECORD. Currently, the
* typoutput functions associated with the pseudotypes will work
* okay, but we might have to try harder in future.
*/
getTypeOutputInfo(indexRelation->rd_opcintype[i],
&foutoid, &typisvarlena);
val = OidOutputFunctionCall(foutoid, values[i]);
}
if (i > 0)
appendStringInfoString(&buf, ", ");
appendStringInfoString(&buf, val);
}
appendStringInfoChar(&buf, ')');
return buf.data;
}
Datum
pgrowlocks(PG_FUNCTION_ARGS)
{
FuncCallContext *funcctx;
HeapScanDesc scan;
HeapTuple tuple;
TupleDesc tupdesc;
AttInMetadata *attinmeta;
Datum result;
MyData *mydata;
Relation rel;
if (SRF_IS_FIRSTCALL())
{
text *relname;
RangeVar *relrv;
MemoryContext oldcontext;
AclResult aclresult;
funcctx = SRF_FIRSTCALL_INIT();
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
/* Build a tuple descriptor for our result type */
if (get_call_result_type(fcinfo, NULL, &tupdesc) != TYPEFUNC_COMPOSITE)
elog(ERROR, "return type must be a row type");
attinmeta = TupleDescGetAttInMetadata(tupdesc);
funcctx->attinmeta = attinmeta;
relname = PG_GETARG_TEXT_P(0);
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
rel = heap_openrv(relrv, AccessShareLock);
/* check permissions: must have SELECT on table */
aclresult = pg_class_aclcheck(RelationGetRelid(rel), GetUserId(),
ACL_SELECT);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_CLASS,
RelationGetRelationName(rel));
scan = heap_beginscan(rel, GetActiveSnapshot(), 0, NULL);
mydata = palloc(sizeof(*mydata));
mydata->rel = rel;
mydata->scan = scan;
mydata->ncolumns = tupdesc->natts;
funcctx->user_fctx = mydata;
MemoryContextSwitchTo(oldcontext);
}
funcctx = SRF_PERCALL_SETUP();
attinmeta = funcctx->attinmeta;
mydata = (MyData *) funcctx->user_fctx;
scan = mydata->scan;
/* scan the relation */
while ((tuple = heap_getnext(scan, ForwardScanDirection)) != NULL)
{
HTSU_Result htsu;
TransactionId xmax;
uint16 infomask;
/* must hold a buffer lock to call HeapTupleSatisfiesUpdate */
LockBuffer(scan->rs_cbuf, BUFFER_LOCK_SHARE);
htsu = HeapTupleSatisfiesUpdate(tuple,
GetCurrentCommandId(false),
scan->rs_cbuf);
xmax = HeapTupleHeaderGetRawXmax(tuple->t_data);
infomask = tuple->t_data->t_infomask;
/*
* a tuple is locked if HTSU returns BeingUpdated, and if it returns
* MayBeUpdated but the Xmax is valid and pointing at us.
*/
if (htsu == HeapTupleBeingUpdated ||
(htsu == HeapTupleMayBeUpdated &&
!(infomask & HEAP_XMAX_INVALID) &&
!(infomask & HEAP_XMAX_IS_MULTI) &&
(xmax == GetCurrentTransactionIdIfAny())))
{
char **values;
values = (char **) palloc(mydata->ncolumns * sizeof(char *));
values[Atnum_tid] = (char *) DirectFunctionCall1(tidout,
PointerGetDatum(&tuple->t_self));
values[Atnum_xmax] = palloc(NCHARS * sizeof(char));
snprintf(values[Atnum_xmax], NCHARS, "%d", xmax);
if (infomask & HEAP_XMAX_IS_MULTI)
{
MultiXactMember *members;
int nmembers;
bool first = true;
bool allow_old;
values[Atnum_ismulti] = pstrdup("true");
allow_old = !(infomask & HEAP_LOCK_MASK) &&
(infomask & HEAP_XMAX_LOCK_ONLY);
nmembers = GetMultiXactIdMembers(xmax, &members, allow_old,
false);
if (nmembers == -1)
{
values[Atnum_xids] = "{0}";
values[Atnum_modes] = "{transient upgrade status}";
values[Atnum_pids] = "{0}";
}
else
{
int j;
values[Atnum_xids] = palloc(NCHARS * nmembers);
values[Atnum_modes] = palloc(NCHARS * nmembers);
values[Atnum_pids] = palloc(NCHARS * nmembers);
strcpy(values[Atnum_xids], "{");
strcpy(values[Atnum_modes], "{");
strcpy(values[Atnum_pids], "{");
for (j = 0; j < nmembers; j++)
{
char buf[NCHARS];
if (!first)
{
strcat(values[Atnum_xids], ",");
strcat(values[Atnum_modes], ",");
strcat(values[Atnum_pids], ",");
}
snprintf(buf, NCHARS, "%d", members[j].xid);
strcat(values[Atnum_xids], buf);
switch (members[j].status)
{
case MultiXactStatusUpdate:
snprintf(buf, NCHARS, "Update");
break;
case MultiXactStatusNoKeyUpdate:
snprintf(buf, NCHARS, "No Key Update");
break;
case MultiXactStatusForUpdate:
snprintf(buf, NCHARS, "For Update");
break;
case MultiXactStatusForNoKeyUpdate:
snprintf(buf, NCHARS, "For No Key Update");
break;
case MultiXactStatusForShare:
snprintf(buf, NCHARS, "Share");
break;
case MultiXactStatusForKeyShare:
snprintf(buf, NCHARS, "Key Share");
break;
}
strcat(values[Atnum_modes], buf);
snprintf(buf, NCHARS, "%d",
BackendXidGetPid(members[j].xid));
strcat(values[Atnum_pids], buf);
first = false;
}
strcat(values[Atnum_xids], "}");
strcat(values[Atnum_modes], "}");
strcat(values[Atnum_pids], "}");
}
}
else
{
values[Atnum_ismulti] = pstrdup("false");
values[Atnum_xids] = palloc(NCHARS * sizeof(char));
snprintf(values[Atnum_xids], NCHARS, "{%d}", xmax);
values[Atnum_modes] = palloc(NCHARS);
if (infomask & HEAP_XMAX_LOCK_ONLY)
{
if (HEAP_XMAX_IS_SHR_LOCKED(infomask))
snprintf(values[Atnum_modes], NCHARS, "{For Share}");
else if (HEAP_XMAX_IS_KEYSHR_LOCKED(infomask))
snprintf(values[Atnum_modes], NCHARS, "{For Key Share}");
else if (HEAP_XMAX_IS_EXCL_LOCKED(infomask))
{
if (tuple->t_data->t_infomask2 & HEAP_KEYS_UPDATED)
snprintf(values[Atnum_modes], NCHARS, "{For Update}");
else
snprintf(values[Atnum_modes], NCHARS, "{For No Key Update}");
}
else
/* neither keyshare nor exclusive bit it set */
snprintf(values[Atnum_modes], NCHARS,
"{transient upgrade status}");
}
else
{
if (tuple->t_data->t_infomask2 & HEAP_KEYS_UPDATED)
snprintf(values[Atnum_modes], NCHARS, "{Update}");
else
snprintf(values[Atnum_modes], NCHARS, "{No Key Update}");
}
values[Atnum_pids] = palloc(NCHARS * sizeof(char));
snprintf(values[Atnum_pids], NCHARS, "{%d}",
BackendXidGetPid(xmax));
}
LockBuffer(scan->rs_cbuf, BUFFER_LOCK_UNLOCK);
/* build a tuple */
tuple = BuildTupleFromCStrings(attinmeta, values);
/* make the tuple into a datum */
result = HeapTupleGetDatum(tuple);
/*
* no need to pfree what we allocated; it's on a short-lived
* memory context anyway
*/
SRF_RETURN_NEXT(funcctx, result);
}
else
{
LockBuffer(scan->rs_cbuf, BUFFER_LOCK_UNLOCK);
}
}
heap_endscan(scan);
heap_close(mydata->rel, AccessShareLock);
SRF_RETURN_DONE(funcctx);
}
/*
* Delete error log of the specified relation. This returns true from master
* iif all segments and master find the relation.
*/
Datum
gp_truncate_error_log(PG_FUNCTION_ARGS)
{
text *relname;
char *relname_str;
RangeVar *relrv;
Oid relid;
bool allResults = true;
relname = PG_GETARG_TEXT_P(0);
/*
* Dispatch the work to segments.
*/
if (Gp_role == GP_ROLE_DISPATCH)
{
int i, resultCount = 0;
StringInfoData sql, errbuf;
PGresult **results;
initStringInfo(&sql);
initStringInfo(&errbuf);
appendStringInfo(&sql,
"SELECT pg_catalog.gp_truncate_error_log('%s')",
text_to_cstring(relname));
results = cdbdisp_dispatchRMCommand(sql.data, true, &errbuf,
&resultCount);
if (errbuf.len > 0)
elog(ERROR, "%s", errbuf.data);
Assert(resultCount > 0);
for (i = 0; i < resultCount; i++)
{
Datum value;
bool isnull;
if (PQresultStatus(results[i]) != PGRES_TUPLES_OK)
ereport(ERROR,
(errmsg("unexpected result from segment: %d",
PQresultStatus(results[i]))));
value = ResultToDatum(results[i], 0, 0, boolin, &isnull);
allResults &= (!isnull && DatumGetBool(value));
PQclear(results[i]);
}
pfree(errbuf.data);
pfree(sql.data);
}
relname_str = text_to_cstring(relname);
if (strcmp(relname_str, "*.*") == 0)
{
/*
* Only superuser is allowed to delete log files across database.
*/
if (!superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
(errmsg("must be superuser to delete all error log files"))));
ErrorLogDelete(InvalidOid, InvalidOid);
}
else if (strcmp(relname_str, "*") == 0)
{
/*
* Database owner can delete error log files.
*/
if (!pg_database_ownercheck(MyDatabaseId, GetUserId()))
aclcheck_error(ACLCHECK_NOT_OWNER, ACL_KIND_DATABASE,
get_database_name(MyDatabaseId));
ErrorLogDelete(MyDatabaseId, InvalidOid);
}
else
{
AclResult aclresult;
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
relid = RangeVarGetRelid(relrv, true);
/* Return false if the relation does not exist. */
if (!OidIsValid(relid))
PG_RETURN_BOOL(false);
/*
* Allow only the table owner to truncate error log.
*/
aclresult = pg_class_aclcheck(relid, GetUserId(), ACL_TRUNCATE);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_CLASS, relrv->relname);
/* We don't care if this fails or not. */
ErrorLogDelete(MyDatabaseId, relid);
}
/* Return true iif all segments return true. */
PG_RETURN_BOOL(allResults);
}
/*
* gp_read_error_log
*
* Returns set of error log tuples.
*/
Datum
gp_read_error_log(PG_FUNCTION_ARGS)
{
FuncCallContext *funcctx;
ReadErrorLogContext *context;
HeapTuple tuple;
Datum result;
/*
* First call setup
*/
if (SRF_IS_FIRSTCALL())
{
MemoryContext oldcontext;
FILE *fp;
text *relname;
funcctx = SRF_FIRSTCALL_INIT();
relname = PG_GETARG_TEXT_P(0);
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
context = palloc0(sizeof(ReadErrorLogContext));
funcctx->user_fctx = (void *) context;
funcctx->tuple_desc = BlessTupleDesc(GetErrorTupleDesc());
/*
* Though this function is usually executed on segment, we dispatch
* the execution if it happens to be on QD, and combine the results
* into one set.
*/
if (Gp_role == GP_ROLE_DISPATCH)
{
int resultCount = 0;
PGresult **results = NULL;
StringInfoData sql;
StringInfoData errbuf;
int i;
initStringInfo(&sql);
initStringInfo(&errbuf);
/*
* construct SQL
*/
appendStringInfo(&sql,
"SELECT * FROM pg_catalog.gp_read_error_log(%s) ",
quote_literal_internal(text_to_cstring(relname)));
results = cdbdisp_dispatchRMCommand(sql.data, true, &errbuf,
&resultCount);
if (errbuf.len > 0)
elog(ERROR, "%s", errbuf.data);
Assert(resultCount > 0);
for (i = 0; i < resultCount; i++)
{
if (PQresultStatus(results[i]) != PGRES_TUPLES_OK)
elog(ERROR, "unexpected result from segment: %d",
PQresultStatus(results[i]));
context->numTuples += PQntuples(results[i]);
}
pfree(errbuf.data);
pfree(sql.data);
context->segResults = results;
context->numSegResults = resultCount;
}
else
{
/*
* In QE, read the error log.
*/
RangeVar *relrv;
Oid relid;
relrv = makeRangeVarFromNameList(textToQualifiedNameList(relname));
relid = RangeVarGetRelid(relrv, true);
/*
* If the relation has gone, silently return no tuples.
*/
if (OidIsValid(relid))
{
AclResult aclresult;
/*
* Requires SELECT priv to read error log.
*/
aclresult = pg_class_aclcheck(relid, GetUserId(), ACL_SELECT);
if (aclresult != ACLCHECK_OK)
aclcheck_error(aclresult, ACL_KIND_CLASS, relrv->relname);
ErrorLogFileName(context->filename, MyDatabaseId, relid);
fp = AllocateFile(context->filename, "r");
context->fp = fp;
}
}
MemoryContextSwitchTo(oldcontext);
if (Gp_role != GP_ROLE_DISPATCH && !context->fp)
{
pfree(context);
SRF_RETURN_DONE(funcctx);
}
}
funcctx = SRF_PERCALL_SETUP();
context = (ReadErrorLogContext *) funcctx->user_fctx;
/*
* Read error log, probably on segments. We don't check Gp_role, however,
* in case master also wants to read the file.
*/
if (context->fp)
{
pg_crc32 crc, written_crc;
tuple = ErrorLogRead(context->fp, &written_crc);
/*
* CRC check.
*/
if (HeapTupleIsValid(tuple))
{
INIT_CRC32C(crc);
COMP_CRC32C(crc, tuple->t_data, tuple->t_len);
FIN_CRC32C(crc);
if (!EQ_CRC32C(crc, written_crc))
{
elog(LOG, "incorrect checksum in error log %s",
context->filename);
tuple = NULL;
}
}
/*
* If we found a valid tuple, return it. Otherwise, fall through
* in the DONE routine.
*/
if (HeapTupleIsValid(tuple))
{
/*
* We need to set typmod for the executor to understand
* its type we just blessed.
*/
HeapTupleHeaderSetTypMod(tuple->t_data,
funcctx->tuple_desc->tdtypmod);
result = HeapTupleGetDatum(tuple);
SRF_RETURN_NEXT(funcctx, result);
}
}
/*
* If we got results from dispatch, return all the tuples.
*/
while (context->currentResult < context->numSegResults)
{
Datum values[NUM_ERRORTABLE_ATTR];
bool isnull[NUM_ERRORTABLE_ATTR];
PGresult *segres = context->segResults[context->currentResult];
int row = context->currentRow;
if (row >= PQntuples(segres))
{
context->currentRow = 0;
context->currentResult++;
continue;
}
context->currentRow++;
MemSet(isnull, false, sizeof(isnull));
values[0] = ResultToDatum(segres, row, 0, timestamptz_in, &isnull[0]);
values[1] = ResultToDatum(segres, row, 1, textin, &isnull[1]);
values[2] = ResultToDatum(segres, row, 2, textin, &isnull[2]);
values[3] = ResultToDatum(segres, row, 3, int4in, &isnull[3]);
values[4] = ResultToDatum(segres, row, 4, int4in, &isnull[4]);
values[5] = ResultToDatum(segres, row, 5, textin, &isnull[5]);
values[6] = ResultToDatum(segres, row, 6, textin, &isnull[6]);
values[7] = ResultToDatum(segres, row, 7, byteain, &isnull[7]);
tuple = heap_form_tuple(funcctx->tuple_desc, values, isnull);
result = HeapTupleGetDatum(tuple);
SRF_RETURN_NEXT(funcctx, result);
}
if (context->segResults != NULL)
{
int i;
for (i = 0; i < context->numSegResults; i++)
PQclear(context->segResults[i]);
/* XXX: better to copy to palloc'ed area */
free(context->segResults[i]);
}
/*
* Close the file, if we have opened it.
*/
if (context->fp != NULL)
{
FreeFile(context->fp);
context->fp = NULL;
}
SRF_RETURN_DONE(funcctx);
}