- Joined
- May 19, 2004
- Messages
- 13,598
- Reaction score
- 7,761
- Location
- Texas
- Gender
- Male
- Political Leaning
- Libertarian - Right
Short version: The ISP was painting their datacenter and the painter accidentally 'elbowed' the emergency power off switch. Databases do not like instant power off. It needed to be repaired. We lost the posts while the server was confused.
Long technical version: The above statement, plus... During the data recovery the server had mixed tokens and session ID's in the system after reboot. That enabled a few to be able to log in and possibly post. However, I was unable to turn off the forum software as the switch is in the database. We were also unable to view/update via phpmyadmin. During this time, the troubleshooters attempted repair from daily and weekly backups and were unsuccessful (due to the errors stated previously). Command line on the system appeared great resulting in the technicians having quick judgments to the current status. When attempting to contact the ports for the db via 80, it would go nuts. The errors (5 of them) combined with our multiple cache system made it even more of a challenge to find the root cause.
Results: The database was eventually recovered from the latest daily backup successfully. Posts after 2:30am Saturday are lost.
Please accept our apology and thank you for your patience during the last two days.
Long technical version: The above statement, plus... During the data recovery the server had mixed tokens and session ID's in the system after reboot. That enabled a few to be able to log in and possibly post. However, I was unable to turn off the forum software as the switch is in the database. We were also unable to view/update via phpmyadmin. During this time, the troubleshooters attempted repair from daily and weekly backups and were unsuccessful (due to the errors stated previously). Command line on the system appeared great resulting in the technicians having quick judgments to the current status. When attempting to contact the ports for the db via 80, it would go nuts. The errors (5 of them) combined with our multiple cache system made it even more of a challenge to find the root cause.
Results: The database was eventually recovered from the latest daily backup successfully. Posts after 2:30am Saturday are lost.
Please accept our apology and thank you for your patience during the last two days.
Last edited: