| From: | Thomas Munro <thomas(dot)munro(at)enterprisedb(dot)com> |
|---|---|
| To: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
| Cc: | Fujii Masao <masao(dot)fujii(at)gmail(dot)com>, Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com>, Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Measuring replay lag |
| Date: | 2017-01-03 23:40:59 |
| Message-ID: | CAEepm=0BWVeDh+0GmhHoZKXwwfW2RGMv95Z4UM=3QW-8=kw0EA@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, Jan 4, 2017 at 12:22 PM, Thomas Munro
<thomas(dot)munro(at)enterprisedb(dot)com> wrote:
> (replay_lag - (write_lag / 2) may be a cheap proxy
> for a lag time that doesn't include the return network leg, and still
> doesn't introduce clock difference error)
(Upon reflection it's a terrible proxy for that because of the mix of
write/flush work done by WAL receiver today, but would improve
dramatically if the WAL writer were doing the flushing. A better yet
proxy might involve also tracking receive_lag which doesn't include
the write() syscall. My real point is that there are ways to work
backwards from the two-way round trip time to get other estimates, but
no good ways to undo the damage that would be done to the data if we
started using two systems' clocks.)
--
Thomas Munro
http://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Lewis, Ian (Microstar Laboratories) | 2017-01-03 23:45:10 | Re: Cluster wide option to control symbol case folding |
| Previous Message | Thomas Munro | 2017-01-03 23:27:11 | Re: Measuring replay lag |