Skip to Content
1

why would recovery urgency suddenly start to grow faster than checkpoint urgency

Feb 10 at 04:40 PM

74

avatar image
Former Member

I have an office using sql anywhere 16, db is 83gigs, using a log file. the recovery urgency is growing much faster than checkpoint causing the recovery urgency to get to 99% often throughout the day. this is causing a flush of data from the log file to the disk. the unscheduled requests will go from 0 to has high has 850 while the flush is happening. all the user hang during this time period and when it is done the unscheduled request go to 0 and all the blocked connections clear.

SQL | help
10 |10000 characters needed characters left characters exceeded
* Please Login or Register to Answer, Follow or Comment.

3 Answers

Breck Carter Feb 25 at 03:40 PM
1

> have an office using sql anywhere 16, db is 83gigs, using a log file.

What build of SQL Anywhere are you using?

> the recovery urgency is growing much faster than checkpoint causing the

> recovery urgency to get to 99% often throughout the day.

Recovery urgency often rises faster than checkpoint urgency. It can also rise above 100% without causing performance problems; 1000% is not unheard of.

> this is causing a flush of data from the log file to the disk.

How do you know "a flush of data from the log file to the disk" is happening?

How do you know checkpoint urgency is causing it?

Checkpoints don't have anything to do with the transaction log file, they have to do with flushing the checkpoint log to the database file. ...or are you talking about a checkpoint happening?

> the unscheduled requests will go from 0 to has high has 850 while the flush is happening.

How many busy connections do you have running at one time?

Are you absolutely sure the unscheduled request rises as high as 850? ...that is an absolutely astounding number, larger than I have ever seen.

> all the user hang during this time period and when it is done the unscheduled

> request go to 0 and all the blocked connections clear.

Are you saying that checkpoints are happening too often, are taking too long, and seem to be causing performance problems?

Share
10 |10000 characters needed characters left characters exceeded
Volker Barth Feb 13 at 11:19 AM
0

Have you change the recovery_time option lately? Or do you run more (or more complex) transactions per minute than before?

Share
10 |10000 characters needed characters left characters exceeded
avatar image
Former Member Feb 16 at 12:34 PM
0

i have not change the recovery_time, the default is in place. I will ck on the transactions

Share
10 |10000 characters needed characters left characters exceeded