Hi all.
We setup a separate location for virtuoso backup files on a different
volume. First I did a full backup by stopping all processes writing to
virtuoso and executed commands in ISQL terminal:
SQL> checkpoint_interval(-1);
SQL> backup_context_clear();
SQL> checkpoint;
SQL>
backup_online('virt-inc_dump_17082106#',1000000,0,vector('/mnt/backups/virtuoso'));
SQL> checkpoint_interval(60);
So I disabled checkpoints while doing backups and enabled after backups
finished. This worked fine.
Then I setup a cron job to execute sql file with same commands at
midnight every day. Not using virtuoso built in solution as needed to
specify different location for backup files, by default they are created
to the virtuoso work directory.
The midnight backup failed right before end with page checksum error. At
midnight the processes writing to virtuoso were running, when the backup
started. And virtuoso service stopped after that error. The virtuoso.log
about the error:
21:00:38 Starting online backup from page 1 to 16455680, current log is:
/var/lib/virtuoso-opensource-7/db/virtuoso.trx
21:00:39 Backing up page 10000
21:00:39 Backing up page 20000
...
21:04:04 Backing up page 16450000
21:04:05 page set checksum ck=2ace7ae3 pck=2aca7ae3 xor = 40000
21:04:05 /usr/bin/virtuoso-t() [0xb07467]
21:04:05 /usr/bin/virtuoso-t() [0xb0750f]
21:04:05 /usr/bin/virtuoso-t() [0x4a2bd1]
21:04:05 /usr/bin/virtuoso-t() [0x4a87ef]
21:04:05 /usr/bin/virtuoso-t() [0x4a8cd8]
21:04:05 /usr/bin/virtuoso-t() [0x4bc9bb]
21:04:05 /usr/bin/virtuoso-t() [0x593e12]
21:04:05 /usr/bin/virtuoso-t() [0x4fd830]
21:04:05 /usr/bin/virtuoso-t() [0x7456c7]
21:04:05 /usr/bin/virtuoso-t() [0x746185]
21:04:05 /usr/bin/virtuoso-t() [0x746b15]
21:04:05 /usr/bin/virtuoso-t() [0x746c06]
21:04:05 /usr/bin/virtuoso-t() [0x747cc7]
21:04:05 /usr/bin/virtuoso-t() [0x747e0e]
21:04:05 /usr/bin/virtuoso-t() [0x747f86]
21:04:05 /usr/bin/virtuoso-t() [0x6515b8]
21:04:05 /usr/bin/virtuoso-t() [0x65a055]
21:04:05 /usr/bin/virtuoso-t() [0x69c18b]
21:04:05 /usr/bin/virtuoso-t() [0x6aaecb]
21:04:05 /usr/bin/virtuoso-t() [0x652864]
21:04:05 /usr/bin/virtuoso-t() [0x65a01e]
21:04:05 /usr/bin/virtuoso-t() [0x69c18b]
21:04:05 /usr/bin/virtuoso-t() [0x6aaecb]
21:04:05 /usr/bin/virtuoso-t() [0x652864]
21:04:05 /usr/bin/virtuoso-t() [0x653d36]
21:04:05 /usr/bin/virtuoso-t() [0x65f822]
21:04:05 /usr/bin/virtuoso-t() [0x69c29f]
21:04:05 /usr/bin/virtuoso-t() [0x69c33d]
21:04:05 /usr/bin/virtuoso-t() [0x6edcb8]
21:04:05 /usr/bin/virtuoso-t() [0x590c1e]
21:04:05 /usr/bin/virtuoso-t() [0x69c1a4]
21:04:05 /usr/bin/virtuoso-t() [0x6a6264]
21:04:05 /usr/bin/virtuoso-t() [0x69c1a4]
21:04:05 /usr/bin/virtuoso-t() [0x69c455]
21:04:05 /usr/bin/virtuoso-t() [0x6ed7b4]
21:04:05 /usr/bin/virtuoso-t() [0x59319c]
21:04:05 /usr/bin/virtuoso-t() [0x69c1a4]
21:04:05 /usr/bin/virtuoso-t() [0x6550f1]
21:04:05 /usr/bin/virtuoso-t() [0x655321]
21:04:05 /usr/bin/virtuoso-t() [0x65a085]
21:04:05 /usr/bin/virtuoso-t() [0x69c18b]
21:04:05 /usr/bin/virtuoso-t() [0x6a9007]
21:04:05 /usr/bin/virtuoso-t() [0x6b2f3e]
21:04:05 /usr/bin/virtuoso-t() [0x6b3375]
21:04:05 /usr/bin/virtuoso-t() [0xb097dc]
21:04:05 /usr/bin/virtuoso-t() [0xb15a4c]
21:04:05 /lib/x86_64-linux-gnu/libpthread.so.0(+0x8182) [0x7f5472cf6182]
21:04:05 /lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7f547230047d]
21:04:05 GPF: disk.c:298 page set checksum error
Did a full backup today again with processes writing to virtuoso stopped
and it passed successfully.
What can cause this? Is some other session/connection causing checkpoint
or changing the page while backup is running? Am I missing some command
that virtuoso built in incremental backup is executing to disable
changes to pages meant for backup?
Sincerely.
Piret
------------------------------------------------------------------------------
_______________________________________________
Virtuoso-users mailing list
Virtuoso-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/virtuoso-users
