Don't use _mdfd_getseg() in mdsyncfiletag().

_mdfd_getseg() opens all segments up to the requested one.  That
causes problems for mdsyncfiletag(), if mdunlinkfork() has
already unlinked other segment files.  Open the file we want
directly by name instead, if it's not already open.

The consequence of this bug was a rare panic in the checkpointer,
made more likely if you saturated the sync request queue so that
the SYNC_FORGET_REQUEST messages for a given relation were more
likely to be absorbed in separate cycles by the checkpointer.

Back-patch to 12.  Defect in commit 3eb77eba.

Author: Thomas Munro
Reported-by: Justin Pryzby
Discussion: https://postgr.es/m/20191119115759.GI30362%40telsasoft.com
This commit is contained in:
Thomas Munro 2019-12-14 15:54:31 +13:00
parent 70c4f500ea
commit c3dc0cdd6b
1 changed files with 39 additions and 16 deletions

View File

@ -1258,25 +1258,48 @@ int
mdsyncfiletag(const FileTag *ftag, char *path)
{
SMgrRelation reln = smgropen(ftag->rnode, InvalidBackendId);
MdfdVec *v;
char *p;
int fd,
result,
save_errno;
bool need_to_close;
/* Provide the path for informational messages. */
p = _mdfd_segpath(reln, ftag->forknum, ftag->segno);
strlcpy(path, p, MAXPGPATH);
pfree(p);
/* See if we already have the file open, or need to open it. */
if (ftag->segno < reln->md_num_open_segs[ftag->forknum])
{
File file;
/* Try to open the requested segment. */
v = _mdfd_getseg(reln,
ftag->forknum,
ftag->segno * (BlockNumber) RELSEG_SIZE,
false,
EXTENSION_RETURN_NULL | EXTENSION_DONT_CHECK_SIZE);
if (v == NULL)
return -1;
file = reln->md_seg_fds[ftag->forknum][ftag->segno].mdfd_vfd;
strlcpy(path, FilePathName(file), MAXPGPATH);
fd = FileGetRawDesc(file);
need_to_close = false;
}
else
{
char *p;
/* Try to fsync the file. */
return FileSync(v->mdfd_vfd, WAIT_EVENT_DATA_FILE_SYNC);
p = _mdfd_segpath(reln, ftag->forknum, ftag->segno);
strlcpy(path, p, MAXPGPATH);
pfree(p);
fd = OpenTransientFile(path, O_RDWR);
if (fd < 0)
return -1;
need_to_close = true;
}
/* Sync the file. */
pgstat_report_wait_start(WAIT_EVENT_DATA_FILE_SYNC);
result = pg_fsync(fd);
save_errno = errno;
pgstat_report_wait_end();
if (need_to_close && CloseTransientFile(fd) != 0)
ereport(WARNING,
(errcode_for_file_access(),
errmsg("could not close file \"%s\": %m", path)));
errno = save_errno;
return result;
}
/*