Even though hourly cron run is consistently reported as completed, watchdog reports that the database is not being restored consistently every hour. Successful db restores are randomly reported either as every hour or every other hour with no noticeable pattern.
Both the Dump settings site reset time interval and cron are set to run every hour.
Within the module readme the following statement, Note: This requires cron to run at least within this interval, makes it sound like cron needs to run at an interval less than the interval setting or am I reading too much into this statement? In other words if reset time is one hour does cron need to be less than an hour?
Or could this inconsistency just be due to the database not following through with the restore when cron runs? Any way to check for this?
Other insight?
| Comment | File | Size | Author |
|---|---|---|---|
| #3 | demo-DRUPAL-5.cron_.patch | 784 bytes | sun |
Comments
Comment #1
sunIf you are only referring to the watchdog table, then the answer is: the watchdog table is not reset when the site is reset. Also, all cache tables are only truncated and not even dumped to reduce dumps in filesize.
If cron runs every hour and you configured Demo to reset the site every hour, the site should be reset every hour.
Comment #2
wiredescape commentedSo if watchdog is untouched we can assume what it is reporting is accurate regardless of whether or not the db restore takes place or not. In this case why would watchdog show that the db is restored sometimes every hour and sometimes every two hours? It seems to be showing that it skips every other cycle.
Your correct that Demo "should" reset every hour but that is not what is being reported...
Comment #3
sunAttached patch should fix this issue - please test.
Comment #4
wiredescape commentedThanks for looking into this Sun.
Unfortunately patch does not solve the problem.
Manually added patch to Official Release 5.x-1.4.
I tested and verified that the DB is not being refreshed when watchdog does not report the db refresh. I tried varying Demo reset times and cron run times with no difference. Over the last +12 hours Watchdog reports that cron consistently runs successfully each time but database refresh is still inconsistent and random.
Unfortunately I do not have access to db logs to check what is happening on that side.
Here is a snippet of watchdog. Times where DB restore does not happen are highlighted:
Type Date Message User
cron 2008-12-01 10:40 Cron run completed. Visitor
demo 2008-12-01 10:40 Successfully restored database from ... Visitor
cron 2008-12-01 10:10 Cron run completed. Visitor
cron 2008-12-01 09:40 Cron run completed. Visitor
demo 2008-12-01 09:40 Successfully restored database from ... Visitor
cron 2008-12-01 09:10 Cron run completed. Visitor
cron 2008-12-01 08:40 Cron run completed. Visitor
demo 2008-12-01 08:40 Successfully restored database from ... Visitor
cron 2008-12-01 08:10 Cron run completed. Visitor
cron 2008-12-01 07:40 Cron run completed. Visitor
demo 2008-12-01 07:40 Successfully restored database from ... Visitor
cron 2008-12-01 07:10 Cron run completed. Visitor
demo 2008-12-01 07:10 Successfully restored database from ... Visitor
cron 2008-12-01 06:40 Cron run completed. Visitor
cron 2008-12-01 06:10 Cron run completed. Visitor
demo 2008-12-01 06:10 Successfully restored database from ... Visitor
cron 2008-12-01 05:40 Cron run completed. Visitor
demo 2008-12-01 05:40 Successfully restored database from ... Visitor
cron 2008-12-01 05:10 Cron run completed. Visitor
demo 2008-12-01 05:10 Successfully restored database from ... Visitor
cron 2008-12-01 04:40 Cron run completed. Visitor
cron 2008-12-01 04:10 Cron run completed. Visitor
demo 2008-12-01 04:10 Successfully restored database from ... Visitor
cron 2008-12-01 03:40 Cron run completed. Visitor
Comment #5
sunComment #6
sun#424524: demo_reset() takes more than 1 second, impacts demo_reset_interval check