Skip site navigation (1) Skip section navigation (2)

Re: PATCH: pg_basebackup (missing exit on error)

From: Fujii Masao <masao(dot)fujii(at)gmail(dot)com>
To: Robert Haas <robertmhaas(at)gmail(dot)com>
Cc: Thomas Ogrisegg <tom-nic(at)patches(dot)fnord(dot)at>, pgsql-hackers(at)postgresql(dot)org
Subject: Re: PATCH: pg_basebackup (missing exit on error)
Date: 2012-03-27 11:13:34
Message-ID: CAHGQGwHXzLFg9KBC=MWaGfVQwh9H_oo9NPnh6VLpy4Js0-MEuw@mail.gmail.com (view raw or flat)
Thread:
Lists: pgsql-hackers
On Thu, Jan 26, 2012 at 11:48 PM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> On Tue, Jan 24, 2012 at 4:39 AM, Thomas Ogrisegg
> <tom-nic(at)patches(dot)fnord(dot)at> wrote:
>> While testing a backup script, I noticed that pg_basebackup exits with
>> 0, even if it had errors while writing the backup to disk when the
>> backup file is to be sent to stdout. The attached patch should fix this
>> problem (and some special cases when the last write fails).
>
> This part looks like a typo:
>
> +                                       if (fflush (tarfile) != 0)
> +                                       {
> +                                               fprintf(stderr, _("%s:
> error flushing stdout: %s\n"),
> +
> strerror (errno));
> +                                               disconnect_and_exit(1);
> +                                       }
>
> The error is in flushing the tarfile, not stdout, right?

No, in this case tarfile is set to stdout as follows.

-------------
		if (strcmp(basedir, "-") == 0)
		{
#ifdef HAVE_LIBZ
			if (compresslevel != 0)
			{
				ztarfile = gzdopen(dup(fileno(stdout)), "wb");
				if (gzsetparams(ztarfile, compresslevel, Z_DEFAULT_STRATEGY) != Z_OK)
				{
					fprintf(stderr, _("%s: could not set compression level %d: %s\n"),
							progname, compresslevel, get_gz_error(ztarfile));
					disconnect_and_exit(1);
				}
			}
			else
#endif
				tarfile = stdout;
		}
-------------

I don't think that pg_basebackup really needs to fflush() stdout for
each file. Right?

-------------
 #endif
 				if (tarfile != NULL)
-					fclose(tarfile);
+				{
+					if (fclose(tarfile) != 0)
+					{
+						fprintf(stderr, _("%s: error closing file \"%s\": %s\n"),
+								progname, filename, strerror (errno));
+						disconnect_and_exit(1);
+					}
+				}
-------------

This message doesn't obey the PostgreSQL message style.

It's guaranteed that the tarfile must not be NULL here, so the above check of
tarfile is not required. The remaining code of pg_basebackup relies on this
assumption.

Attached patch removes the fflush() part, changes the log message and removes
the check of tarfile, as above.

Regards,

-- 
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center

Attachment: missing_exit_on_error_in_pg_basebackup_v1.patch
Description: text/x-diff (1.6 KB)

In response to

Responses

pgsql-hackers by date

Next:From: Shigeru HANADADate: 2012-03-27 11:14:09
Subject: Re: pgsql_fdw, FDW for PostgreSQL server
Previous:From: Fujii MasaoDate: 2012-03-27 09:40:34
Subject: Re: Reporting WAL file containing checkpoint's REDO record in pg_controldata's result

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group