Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commit0f12905

Browse files
committed
Count individual SQL commands in pg_restore's --transaction-size mode.
The initial implementation in commit959b38d counted one actionper TOC entry (except for some special cases for multi-blob BLOBSentries). This assumes that TOC entries are all about equallycomplex, but it turns out that that assumption doesn't hold up verywell in binary-upgrade mode. For example, even after the previouscommit I was able to cause backend bloat with tables having manyinherited constraints. There may be other cases too. (Since noserious problems have been reported with --single-transaction mode,we can conclude that the backend copes well with psql's regularrestore scripts; but before959b38d we never ran binary-upgraderestores with multi-command transactions.)To fix, count multi-command TOC entries as N actions, allowing thetransaction size to be scaled down when we hit a complex TOC entry.Rather than add a SQL parser to pg_restore, approximate "multicommand" by counting semicolons in the TOC entry's defn string.This will be fooled by semicolons appearing in string literals ---but the error is in the conservative direction, so it doesn't seemworth working harder. The biggest risk is with function/procedureTOC entries, but we can just explicitly skip those.(This is undoubtedly a hack, and maybe someday we'll be able torevert it after fixing the backend's bloat issues or rethinkingwhat pg_dump emits in binary upgrade mode. But that surely isn'ta project for v17.)Thanks to Alexander Korotkov for the let's-count-semicolons idea.Per report from Justin Pryzby. Back-patch to v17 where txn_size modewas introduced.Discussion:https://postgr.es/m/ZqEND4ZcTDBmcv31@pryzbyj2023
1 parentb3f0e05 commit0f12905

File tree

1 file changed

+25
-3
lines changed

1 file changed

+25
-3
lines changed

‎src/bin/pg_dump/pg_backup_archiver.c

Lines changed: 25 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -3827,10 +3827,32 @@ _printTocEntry(ArchiveHandle *AH, TocEntry *te, bool isData)
38273827
{
38283828
IssueACLPerBlob(AH,te);
38293829
}
3830-
else
3830+
elseif (te->defn&&strlen(te->defn)>0)
38313831
{
3832-
if (te->defn&&strlen(te->defn)>0)
3833-
ahprintf(AH,"%s\n\n",te->defn);
3832+
ahprintf(AH,"%s\n\n",te->defn);
3833+
3834+
/*
3835+
* If the defn string contains multiple SQL commands, txn_size mode
3836+
* should count it as N actions not one. But rather than build a full
3837+
* SQL parser, approximate this by counting semicolons. One case
3838+
* where that tends to be badly fooled is function definitions, so
3839+
* ignore them. (restore_toc_entry will count one action anyway.)
3840+
*/
3841+
if (ropt->txn_size>0&&
3842+
strcmp(te->desc,"FUNCTION")!=0&&
3843+
strcmp(te->desc,"PROCEDURE")!=0)
3844+
{
3845+
constchar*p=te->defn;
3846+
intnsemis=0;
3847+
3848+
while ((p=strchr(p,';'))!=NULL)
3849+
{
3850+
nsemis++;
3851+
p++;
3852+
}
3853+
if (nsemis>1)
3854+
AH->txnCount+=nsemis-1;
3855+
}
38343856
}
38353857

38363858
/*

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp