Backup is failing

[joximu]

Could you post the error message again?

and the output of this sql query:
SELECT sqld_id, sqld_name FROM sql_database

/J

[dirkson]

Re: Haeber
No effect on the error message.

Re: joximu:
Err... Like it says earlier in the thread, "Can't use string ("") as an ARRAY ref while "strict refs" in use at /var/www/ispcp/engine/backup/ispcp-backup-all line 324." This error message is from /var/log/ispcp/ispcp-backup-all-mngr.log
Like the OP, I'm also running Debian, but I'm running Debian Etch (4.0, I think). Also, my version of ISPCP is one back from the currently released version.

Output from SELECT sqld_id, sqld_name FROM sql_database:

Code:

mysql> SELECT sqld_id, sqld_name FROM sql_database;
+---------+-----------------+
| sqld_id | sqld_name       |
+---------+-----------------+
|       1 | 32_fcemail      |
|       2 | 33_glbl         |
|       3 | 37_aplied       |
|       4 | 21_mpcwordpress |
|       5 | 39_lvsd         |
|       6 | 41_passed       |
+---------+-----------------+

That's about right for number of databases. There ARE a LOT of websites that don't have mysql databases. All the databases appear to exist, but #1 (32_fcemail) appears to be empty.

[joximu]

Until which domain the backup works properly?

/J

[dirkson]

It works for the first domain (frogstar.com , doesn't have an entry in the last posted MYSQL data) then stops. I'm not sure how to determine which domain is second in its order.

[joximu]

Order should be this one:

Code:

SELECT
                        t1.domain_id,
                        t1.domain_name,
                        t1.domain_status,
                        t1.domain_uid,
                        t1.domain_gid,
                        t2.admin_name,
                        t2.email
                FROM
                        domain AS t1,
                        admin AS t2
                WHERE
                                t1.domain_status        = 'ok'
                        AND     t1.domain_admin_id      = t2.admin_id
                ORDER BY
                        t1.domain_id

Maybe you can set DEBUG to on/1 in the ispcp.conf and then run the backup script manually...

/J

[dirkson]

Hey, that was useful! Error message, now with verbosity:

Code:

# cat /var/log/ispcp/ispcp-backup-*
DEBUG: push_el() sub_name: backup_all_start_up(), msg: Starting...
DEBUG: push_el() sub_name: lock_backup_all_system(), msg: Starting...
DEBUG: push_el() sub_name: sys_command(), msg: Starting...
DEBUG: push_el() sub_name: sys_command('/usr/bin/touch /tmp/ispcp-backup-all.lock'), msg: Ending...
DEBUG: push_el() sub_name: lock_backup_all_system(), msg: Ending...
DEBUG: push_el() sub_name: get_conf(), msg: Starting...
DEBUG: push_el() sub_name: get_file(), msg: Starting...
DEBUG: push_el() sub_name: get_file(), msg: Ending...
DEBUG: push_el() sub_name: setup_main_vars(), msg: Starting...
DEBUG: push_el() sub_name: decrypt_db_password(), msg: Starting...
DEBUG: push_el() sub_name: decrypt_db_password(), msg: Ending...
DEBUG: push_el() sub_name: setup_main_vars(), msg: Ending...
DEBUG: push_el() sub_name: get_conf(), msg: Ending...
DEBUG: push_el() sub_name: backup_all_start_up(), msg: Ending...
DEBUG: push_el() sub_name: database_revision(), msg: Starting...
DEBUG: push_el() sub_name: doSQL(), msg: Starting...
DEBUG: push_el() sub_name: doSQL(), msg: Ending...
DEBUG: push_el() sub_name: database_revision(), msg: Ending...
DEBUG: push_el() sub_name: backup_all_engine(), msg: Starting...
DEBUG: push_el() sub_name: doSQL(), msg: Starting...
DEBUG: push_el() sub_name: doSQL(), msg: Ending...
DEBUG: push_el() sub_name: get_human_date(), msg: Starting...
DEBUG: push_el() sub_name: get_human_date(), msg: Ending...
DEBUG: push_el() sub_name: doSQL(), msg: Starting...
DEBUG: push_el() sub_name: doSQL(), msg: Ending...
DEBUG: push_el() sub_name: sys_command(), msg: Starting...
DEBUG: push_el() sub_name: sys_command('/bin/tar --create --directory=/var/www/virtual/frogstar.com --gzip --file=/var/www/virtual/frogstar.com/frogstar.com-backup-2009.03.24-101915.tar.gz --exclude=logs --exclude=phptmp --exclude=backups --exclude=frogstar.com-backup-2009.03.24-101915.tar.gz . 2> /var/www/virtual/frogstar.com/frogstar.com-backup-2009.03.24-101915.tar.gz.log'), msg: Ending...
DEBUG: push_el() sub_name: sys_command(), msg: Starting...
DEBUG: push_el() sub_name: sys_command('/bin/rm -rf /var/www/virtual/frogstar.com/frogstar.com-backup-2009.03.24-101915.tar.gz.log'), msg: Ending...
DEBUG: push_el() sub_name: sys_command(), msg: Starting...
DEBUG: push_el() sub_name: sys_command('/bin/rm -rf /var/www/virtual/frogstar.com/backups/frogstar.com-*.tar.*'), msg: Ending...
DEBUG: push_el() sub_name: sys_command(), msg: Starting...
DEBUG: push_el() sub_name: sys_command('/bin/mv -f /var/www/virtual/frogstar.com/frogstar.com-backup-2009.03.24-101915.tar.gz /var/www/virtual/frogstar.com/backups'), msg: Ending...
DEBUG: push_el() sub_name: setfmode(), msg: Starting...
DEBUG: push_el() sub_name: setfmode(), msg: Ending...
DEBUG: push_el() sub_name: get_human_date(), msg: Starting...
DEBUG: push_el() sub_name: get_human_date(), msg: Ending...
DEBUG: push_el() sub_name: doSQL(), msg: Starting...
DEBUG: push_el() sub_name: doSQL(), msg: ERROR: Incorrect SQL Query -> MySQL server has gone away
Can't use string ("") as an ARRAY ref while "strict refs" in use at /var/www/ispcp/engine/backup/ispcp-backup-all line 324.

So mysql has "gone away"? This is odd- mysql never gives any issues elsewhere, and it's being used all the time.

First two entries (All that matter) from the mysql statement:

Code:

+-----------+-------------------------------+---------------+------------+------------+-------------------------------+----------------------+
| domain_id | domain_name                   | domain_status | domain_uid | domain_gid | admin_name                    | email                |
+-----------+-------------------------------+---------------+------------+------------+-------------------------------+----------------------+
|         1 | frogstar.com                  | ok            |       2004 |       2004 | frogstar.com                  | dirkson@frogstar.com |
|         2 | arizer.com                    | ok            |       2011 |       2011 | arizer.com                    | dirkson@frogstar.com |

It appears to be borking on arizer.com, which is only about 50mb. It appears to do frogstar.com just fine, which is 500mb or so. Neither of these currently have a mysql database.

[joximu]

I think it appears after the frogstar file backup when the script tries to find out if frogstar also has mysql tables to be backed up.

Now its time to find out what about this "mysql has gone away..."
See http://dev.mysql.com/doc/refman/5.1/en/gone-away.html

Maybe enable log in mysql and try to find out more...

/J

[dirkson]

THERE we go! That's done it! Reading that little mysql site, I managed to track the issue down- My mysql timeout was set to 30 seconds, rather than the more usual 8 hours. A little bump up to 10 minutes seems to have done the trick, and now ispcp is backing up nicely. I'm going to bet the original poster also had this problem- Gzip is faster at zipping than bzip2, and therefore finished under his timeout restriction. Awesome! Thanks for your help!

[joximu]

Thanx - this might also be important for others...

/J

[Norri]

This helped me a lot! Thanks