I'm performing a backup restore on a 300GB database and it's taking a long time. It's coming up for 4 hours since it stopped creating the extent files, and started populating them with the data. Is there any way I can work out how far it has got in terms of how much longer I can expect it to run?
Proenv says "
OpenEdge Release 11.2 as of Wed Feb 13 19:01:30 EST 2013
Start of extending target DB to needed size... (9432)
This is a full backup of F:\DATABASE\LIVE\icmasliv.db. (6759)
This backup was taken Fri Apr 25 14:49:15 2014. (6760)
The blocksize is 8192. (6994)
It will require a minimum of 47665716 blocks to restore. (6763)
Start of restoring the target DB... (9433)"
The log file says even less.
Thanks in advance.
Look at the time stamps on the extents?
1) Experience is a really good estimator -- just one of the many reasons that you should periodically restore backups.
2) You could look at the IO rate on the target device and then do some math -- a minimum of approximately 48M blocks 8k need to be restored -- if your disk subsystem is currently doing about 1,000 per second that would mean roughly 48,000 seconds in total...
3) Just guessing but this situation sounds like a demonstration of how easy it is to saturate the RAM cache on a SAN and the joyous results of doing so.
Yeah I remember I used to be able to look at the time stamps, Thomas, but for some reason they all were stuck at 16:30.
And thanks for your tips too Tom.
As it happens the process finished just as I hit Submit!
Windows has some very strange ideas about when it is appropriate to update file timestamps.
You don't have to rely on Windows Explorer for that. Run Resource Monitor (perfmon /res) and filter for the _dbutil.exe process. You'll see which extent file(s) it is writing to, as well as throughput, disk queuing, etc.
There's also prorest -verbose.
Thanks Rob - really helpful.
Think I've got some disk performance issues. 6 hours to back up a 300GB database with -REPLTargetCreation :(
I have a client where the nightly backup of the 350Gb DB takes 26 hours. Does anyone see the problem?
I'm sure that the bean counters look at 110% utilization as a good thing -- after all, isn't the point to get the most out of your investment?
James,
Are you backup up to the same physical disk where the database resides? What kind of write speed are you getting in MB/s?
James,
Are you backup up to the same physical disk where the database resides? What kind of write speed are you getting in MB/s?
Flag this post as spam/abuse.
Click here to report this email as spam.