June 10, 2013 at 4:18 pm
Recently my server admin restored a DPM backup to a new DB. DPM has its quirks about restoring a renamed DB to the same file groups as the production server so the admin redirected the mdf and ldf to a different storage group, renamed them and copied them back to the main with the test name. The test DB was attached without incident. All went well until 5 days later when the log space on the test db exceeded the space on the allocated disk. At the time of the restore it was about 300MB and 5 days later with no activity from the requesting user it is 38GB. I was asked to determine why the log could not be shrunk.
Puzzled at why there was so much log created I looked into the usage and found none for the DB. We attempted to backup and restore the db using normal utilities and all was successful but the log still persisted. Looking deeper I found replication articles for the test db in the replication folder or the server. The db it was restored from had replication but I was under the understanding that a backup would not carry the replication articles. When the articles were deleted the log disk returned to 300MB but the data tables inflated by about 35GB. We then shrank the data tables with the reclaim option and the data space returned to normal.
My question, why did the replication articles get picked up from the backup, why did they restore and attempt to activate replication generating logs, why did the log data migrate to the data tables and then disappear in an instant of time it took to shrink the 8GB db dataspace.
Steve D
Viewing 0 posts
You must be logged in to reply to this topic. Login to reply