| From: | Don Seiler <don(at)seiler(dot)us> |
|---|---|
| To: | Rui DeSousa <rui(dot)desousa(at)icloud(dot)com> |
| Cc: | Johannes Truschnigg <johannes(at)truschnigg(dot)info>, pgsql-admin <pgsql-admin(at)postgresql(dot)org> |
| Subject: | Re: Streaming Replication Networking Best Practices? |
| Date: | 2018-05-14 20:22:08 |
| Message-ID: | CAHJZqBATEKNrj2J6SZ-kdJr7zPeq4aE36qLcwMUEpia7Nsk1=Q@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-admin |
On Mon, May 14, 2018 at 2:53 PM, Rui DeSousa <rui(dot)desousa(at)icloud(dot)com> wrote:
> What is the latency between the two sites? That is going to dictate how
> much potential throughput a given TCP/IP stream could use. You may need
> tune your TCP/IP window to increase the throughput.
>
Ping shows a pretty solid latency of 49-50 ms.
47 packets transmitted, 47 received, 0% packet loss, time 46057ms
rtt min/avg/max/mdev = 49.125/49.260/49.682/0.289 ms
> I would check the network to ensure there are no errors occurring; i.e.
> dropped packets and retransmits.
>
Looking at ifconfig, I do see dropped RX packets on the iface that talks to
the replica. I don't see any dropped packets on the replica iface that
talks to the primary. The local LAN iface for both also has many dropped
packets.
> Do you have QOS enabled and if so how is that handled? Does it just drop
> the packets? As that would cause the TCP/IP to backoff and retransmit.
>
I'll follow-up with my sys admin team here. I believe there is QoS enabled
but don't have details yet.
Don.
--
Don Seiler
www.seiler.us
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Peter Eisentraut | 2018-05-14 20:33:36 | Re: Streaming Replication Networking Best Practices? |
| Previous Message | Rui DeSousa | 2018-05-14 19:53:54 | Re: Streaming Replication Networking Best Practices? |