|Project:||Backup and Migrate|
I have one site where the restore process takes a fair amount of time. It completes eventually, but not before Apache times out waiting for the PHP child to return a result. Basically, unless Apache gets some feedback from PHP within the limit set by its timeout parameter, it assumes PHP has hung. I'm getting " (70007)The timeout specified has expired: ap_content_length_filter: apr_bucket_read() failed" in my /usr/local/apache/logs/error_log the exact number of seconds after the process began that my Apache timeout would lead me to expect.
Is there any way that the module could use an Ajax call-back to report its progress on a restore to prevent this? Perhaps reconfigure it as a batch job, similar to when a number of database update hook functions need calling from update.php. I appreciate that the transactional nature of a restore is probably the issue here, but some kind of mitigation would be helpful.