Обсуждение: UDP Overflow / UDP Drops on Standby Postgres Service
Postgres General List,
I am stumped trying to prevent an overflowing UDP buffer on a standby Postgres
service. Any help would be most appreciated.
~~~~~~~~
Essentially a UDP buffer associated with the pg_standby process on my localhost
interface gradually fills up once I start Postgres until it hits its maximum
capacity and then proceeds to steadily drop packets. A restart of Postgres (of
course) clears the buffer, but then it begins filling up again.
As far as I can tell, this is not actually causing any problems. (It is only
happening to the standby service, and failover data recovery shows nothing
missing.) Nevertheless, I don't want any buffers to overflow.
(I have also posted this question to ServerFault (http://serverfault.com/questions/564905/udp-overflow-udp-drops-on-standby-postgres-service). That posting has even more detail than I have provided below, such as how I identified pg_standby by querying the /proc files.)
~~~~~~~~
==Salient points==:
a) by querying "/proc" information for UDP I can see non-empty buffers, and identify the "pg_standby" process as the culprit
b) the overflow occurs even when my firewalls on both servers (iptables) are shut down
c) my UDP buffers at 16MB+ seem more than big enough. I could make them larger but that would only mask the problem
d) online discussions of similar problems seem to finger either older versions of Postgres or the Statistics Collector; to rule this out I have tried to turn off all statistics collection (track_activites/counts = off), but the problem continues:
e) a verbose wire sniff of the UDP packet shows nothing useful
f) there is not a great deal of database activity (e.g. roughly one 16MB WAL file is replicated from the primary to the secondary service every 45 minutes)
g) I formerly ran Postgres 8.3.5, with an otherwise identical setup; this problem only began when I upgraded to 9.1.9
~~~~~~~~
==Background on my setup==:
-- two CentOS 6.4 x86_64 bit systems (VMs), each running Postgres 9.1.9, each in a geographically separated (<50 miles) datacenter
-- Postgres is active on my primary server and running in standby mode on my backup:
---- the backup Postgres service is receiving its data two ways:
------ as a warm standby processing WAL files via log shipping
------ on failover the current WAL file on the primary (not yet shipped) is recovered from a DRBD partition synced from the primary box
-- nothing else (of consequence) runs on these boxes except Postgres
~~~~~~~~
Thanks,
Daniel
Postgres Question <postgresquestion@gmail.com> writes:
> Essentially a UDP buffer associated with the pg_standby process on my
> localhost interface gradually fills up once I start Postgres until it
> hits its maximum capacity and then proceeds to steadily drop packets. A
> restart of Postgres (of course) clears the buffer, but then it begins
> filling up again.
AFAICS, there is nothing at all in pg_standby that does IPC of any sort,
let alone UDP transmission in particular. It just does filesystem
access. Perhaps you're running on an NFS filesystem and the UDP traffic
is associated with that? If so, though, any failure to collect packets
would be a bug in glibc, not pg_standby.
A different theory is that pg_standby is just the last surviving child
of some process that opened a UDP port and never closed it. What is
spawning pg_standby in your system?
regards, tom lane