| From: | "Jehan-Guillaume (ioguix) de Rorthais" <ioguix(at)free(dot)fr> |
|---|---|
| To: | Tiemen Ruiten <t(dot)ruiten(at)tech-lab(dot)io> |
| Cc: | pgsql-general(at)lists(dot)postgresql(dot)org |
| Subject: | Re: very high replay_lag on 3-node cluster |
| Date: | 2019-07-22 14:12:28 |
| Message-ID: | 20190722161228.12d0cd03@firost |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Mon, 22 Jul 2019 12:58:47 +0200
Tiemen Ruiten <t(dot)ruiten(at)tech-lab(dot)io> wrote:
[...]
> I've attached a graph of network IO on all servers. The network config is
> identical for all three nodes: 2x bonded gigabit connection to the same
> stacked switch pair.
AFAICS, Network doesn't looks saturated.
> Currently I don't have much metrics on raw disk IO, I can add some
> monitoring items and report those tomorrow if that would be useful.
Yes, ZFS metrics and underlying hardware would be useful. You can add some
detailed CPU and MEM metrics as well, it often helps correlating things
together.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Luca Ferrari | 2019-07-22 19:17:45 | Re: Rearchitecting for storage |
| Previous Message | Rory Campbell-Lange | 2019-07-22 13:06:04 | Re: How to run a task continuously in the background |