Transaction Log backups while Full/Diff backups are running
SQL_ME_RICH
Posts: 112
Hello -
I modified a backup schedule to exclude my TLog backups while my Full backup is running on my main PROD databases. The reason is that I noticed that the only TLogs that were being included (if I for instance needed to do a restore) didn't include anything prior to the full and differential backups completing (Full at 12am and Diff being done at 10am local time).
My question is this...Should I just allow TLogs to go during my Full backup (which takes just under 4 hours, so it starts at midnight local time, and ends shortly before 4am local time), or is it better to have them shut off during that window when the Full is running? I have them going when the Differentials run during the day and evening, but any insight into this would be appreciated.
Many thanks!
I modified a backup schedule to exclude my TLog backups while my Full backup is running on my main PROD databases. The reason is that I noticed that the only TLogs that were being included (if I for instance needed to do a restore) didn't include anything prior to the full and differential backups completing (Full at 12am and Diff being done at 10am local time).
My question is this...Should I just allow TLogs to go during my Full backup (which takes just under 4 hours, so it starts at midnight local time, and ends shortly before 4am local time), or is it better to have them shut off during that window when the Full is running? I have them going when the Differentials run during the day and evening, but any insight into this would be appreciated.
Many thanks!
Comments
There are ways to speed this up.
Chris
The main database is 318gb and growing...The other 3 are about roughly 50gb combined. CHECKSUM and VERIFY are NOT being used for these backups, for the very reason of time constraints. We are able to get around this by doing daily restores of the backups from the night before into our Testing instance.
The main database takes 3hours and 43mins to do a Full backup. The other 3 combined take about an hour. All Fulls and Diffs (as well as TLogs for that matter) run at the same times every day/night.
One of the original benefits of using backup compression was to bring the time taken to run the backup down from what the native maintenance plan performed at.
Have you tried some of the many backup options to bring the backup time down?
Chris
I am not certain to the speed of the platters we are spinning on the internal array for which this db resides (I want to say they are at least 10k, but they could very well be 7200's as well - I doubt they are 15k's, and I am certain that they are not SSD).
I will investigate some of the options of compression and possibly splitting the backup into multiple files, but for now - the time is not a overwhelming concern.
Thanks for sharing your experiences with me though - I know now that I cannot run simultaneous backups of the same db (i.e. fulls/tlog, or diffs/tlog).
Chris
It will show full size on the host side though, but that is what I am being told as to why we are not going to compress them.
Chris
I love what I do! :-D
I share your pain as who would want to be the one to cause a data problem by implementing a new scheme. I have been waiting to implement split backups on our big production database as we cross restore the backups to testing servers and don't want to cause problems.
Chris
I've gotten it down to about a 15/30 minute window of vulnerability, but that is a lot better than the 23 hour and 59 minute window of vulnerability they use to have.
Trust me - there is enough about this beast that has compromises in it that this has become more of a joy to manage than the other parts that - well - off-topic.
:roll:
Your full backup must be 300Gb plus and that first log backup is how big?
Chris
Does that tell me that since it tried to run before the Full had a chance to complete, at 4am, that I did not get the full log that it otherwise should have?
Chris
We ended up ditching the Deduplication Algorithm on our iSCSI controller in favor of compression. The comparison in disk space savings truly is noticeable (and that's an understatement!)
I will have to see what my first Differential looks like come this weekend, but the level 4 compression mode is giving us anywhere between 80%-95% size savings. Pretty awesome, and I thank you for compelling me to look into it more!
Glad to here you are running better now.
Chris