Thread: Status of the startup process post end of recovery
Hello, While doing some tests, I observed that the "ps" command continues to display status of the startup process as "recovering <filename>" even though it has finished the redo and is now performing end-of-recovery checkpoint. This checkpoint can take a while, a few seconds to a few minutes depending on the dirty buffers at the end of the recovery. The "ps" command keep displaying a stale status for all that time. How about calling set_ps_display() with appropriate status string, say similar to what is logged by LogCheckpointStart() ? If we do that, what would be the best place to add that code ? I was thinking adding that to CreateCheckpoint() itself so that the status is displayed at all relevant places. Comments ? Thanks, Pavan -- Pavan Deolasee http://www.linkedin.com/in/pavandeolasee
On 2013-03-04 18:21:50 +0530, Pavan Deolasee wrote: > Hello, > > While doing some tests, I observed that the "ps" command continues to > display status of the startup process as "recovering <filename>" even > though it has finished the redo and is now performing end-of-recovery > checkpoint. This checkpoint can take a while, a few seconds to a few > minutes depending on the dirty buffers at the end of the recovery. The > "ps" command keep displaying a stale status for all that time. How > about calling set_ps_display() with appropriate status string, say > similar to what is logged by LogCheckpointStart() ? If we do that, > what would be the best place to add that code ? I was thinking adding > that to CreateCheckpoint() itself so that the status is displayed at > all relevant places. That might be confusing for example in CREATE DATABASE. Greetings, Andres Freund -- Andres Freund http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services
On Mon, Mar 4, 2013 at 6:23 PM, Andres Freund <andres@2ndquadrant.com> wrote: > On 2013-03-04 18:21:50 +0530, Pavan Deolasee wrote: >> If we do that, >> what would be the best place to add that code ? I was thinking adding >> that to CreateCheckpoint() itself so that the status is displayed at >> all relevant places. > > That might be confusing for example in CREATE DATABASE. > Yeah, you're right. StartupXLOG() seems to be a better place in that case. We can set it before calling CreateCheckpoint() and set it back to "" once checkpoint finishes. Or we can also set it to report the next action such as "running recovery_end_command", but not sure if any step beyond CHECKPOINT would really take long enough to be worth caring for. Thanks, Pavan -- Pavan Deolasee http://www.linkedin.com/in/pavandeolasee