Thread: failures in pg_restore -- possibly corrupt archive
I am attempting to restore a pg dump file and am getting this error message:
pg_restore: [custom archiver] could not find block ID 229750 in archive -- possibly corrupt archive
pg_restore: [custom archiver] could not find block ID 228874 in archive -- possibly corrupt archive
pg_restore: [archiver] worker process failed: exit code 1
The pg_restore command used is:
pg_restore -j 12 -d mydb -h myhost -U username /restore/file.dump
pg_dump was used to created the backup.
I have searched around and have found no information that would lead to a remedy for a successful restore. We have analyzed the disk for bad blocks and could find no problem there. There is no additional information provided than the messages listed above.
Any ideas?
AccessDevelopment TEST <caleb.access@gmail.com> writes: > I am attempting to restore a pg dump file and am getting this error message: > pg_restore: [custom archiver] could not find block ID 229750 in archive > -- possibly corrupt archive > pg_restore: [custom archiver] could not find block ID 228874 in archive > -- possibly corrupt archive > pg_restore: [archiver] worker process failed: exit code 1 > The pg_restore command used is: > pg_restore -j 12 -d mydb -h myhost -U username /restore/file.dump This looks suspiciously like some old complaints about unsupported cases in parallel restore. Does it work if you omit the -j option? If so, what versions of pg_dump and pg_restore are you using exactly, and what did the pg_dump command look like? regards, tom lane
The pg_dump and pg_restore look to be the same version on both boxes ( 9.1.8 ). The pg_dump looks like this: pg_dump \ --host=myhost \ --port=5432 \ --username=myusername \ --format=c \ --file=backup_file.dump \ mydb *Additional notes:* The pg_dump is running on a box with CentOS 6.4 / PostgreSQL 9.1.8 The pg_restore is then executed on a box with CentOS 5.9 / PostgreSQL 9.1.8 The .dump file is 41 GB A $yum list | grep libpq yields nothing on the CentOS 6.4 box but gives this on 5.9 *$ yum list | grep libpq* libpqxx.i386 2.6.8-13.el5epel libpqxx.x86_64 2.6.8-13.el5epel libpqxx-devel.i386 2.6.8-13.el5epel libpqxx-devel.x86_64 2.6.8-13.el5 epel We have not tried without the -j but will give it a go. Please let me know if there are any additional thoughts giving the new info above.
On Thu, Jan 2, 2014 at 3:31 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
AccessDevelopment TEST <caleb.access@gmail.com> writes:This looks suspiciously like some old complaints about unsupported cases
> I am attempting to restore a pg dump file and am getting this error message:
> pg_restore: [custom archiver] could not find block ID 229750 in archive
> -- possibly corrupt archive
> pg_restore: [custom archiver] could not find block ID 228874 in archive
> -- possibly corrupt archive
> pg_restore: [archiver] worker process failed: exit code 1
> The pg_restore command used is:
> pg_restore -j 12 -d mydb -h myhost -U username /restore/file.dump
in parallel restore. Does it work if you omit the -j option? If so,
what versions of pg_dump and pg_restore are you using exactly, and what
did the pg_dump command look like?
regards, tom lane