Re: Postgres Stale Statistics

Hi Adelino,

About the EAGAIN (Resource temporarily unavailable).
> UDP is a stateless protocol, unlike TCP which is connection oriented. The
> recvfrom() code will not know whether or not the sender has closed its
> socket, it only knows whether or not there is data waiting to be read.
> According to the man page for recvfrom on Linux:
> If no messages are available at the socket, the receive calls wait for
> a message to arrive, unless the socket is nonblocking (see fcntl(2)) in
> which case the value -1 is returned and the external variable errno set to
> EAGAIN.

Are you saying the message '-1 EAGAIN (Resource temporarily unavailable)'
is normal?

may you need to explore other option like disk saturation.
> using stale statistics instead of current ones because stats collector is
> not responding
> <https://opensourcedbtech.com/2018/04/03/using-stale-statistics-instead-of-current-ones-because-stats-collector-is-not-responding/>

There is no disk saturation from what we see.

Thanks,
Nikhil

On Thu, Apr 28, 2022 at 2:52 PM Adelino Silva <adelino(dot)silva(at)pt(dot)ibm(dot)com>
wrote:

> Hi Nikhil,
>
> About the EAGAIN (Resource temporarily unavailable).
> UDP is a stateless protocol, unlike TCP which is connection oriented. The
> recvfrom() code will not know whether or not the sender has closed its
> socket, it only knows whether or not there is data waiting to be read.
> According to the man page for recvfrom on Linux:
>
> If no messages are available at the socket, the receive calls wait for
> a message to arrive, unless the socket is nonblocking (see fcntl(2)) in
> which case the value -1 is returned and the external variable errno set to
> EAGAIN.
>
>
> may you need to explore other option like disk saturation.
> using stale statistics instead of current ones because stats collector is
> not responding
> <https://opensourcedbtech.com/2018/04/03/using-stale-statistics-instead-of-current-ones-because-stats-collector-is-not-responding/>
>
> Regards,
>
> Adelino Silva
> ------------------------------
> *From:* Nikhil Shetty <nikhil(dot)dba04(at)gmail(dot)com>
> *Sent:* Wednesday, April 27, 2022 4:36 PM
> *To:* Adelino Silva <adelino(dot)silva(at)pt(dot)ibm(dot)com>
> *Cc:* Pgsql-admin <pgsql-admin(at)lists(dot)postgresql(dot)org>
> *Subject:* [EXTERNAL] Re: Postgres Stale Statistics
>
> Hi Adelino, I went through the article and I see there is no issue with
> IPv6 in our case, it is using IPv4. I used strace and found 'Resource
> temporarily unavailable' error though, not sure what this means, does this
> mean there is an
> ZjQcmQRYFpfptBannerStart
> This Message Is From an External Sender
> This message came from outside your organization.
>
> ZjQcmQRYFpfptBannerEnd
> Hi Adelino,
>
> I went through the article and I see there is no issue with IPv6 in our
> case, it is using IPv4.
>
>
> I used strace and found 'Resource temporarily unavailable' error though,
> not sure what this means, does this mean there is an issue with disk I/O?
>
> strace: Process 5134 attached
>
> epoll_wait(3, [{EPOLLIN, {u32=31860224, u64=31860224}}], 1, -1) = 1
>
> close(3) = 0
>
> recvfrom(10, "\2\0\0\0\230\0\0\0\7(at)\0\0\1\0\0\0\5\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
> 1000, 0, NULL, NULL) = 152
>
> recvfrom(10, 0x7ffeeb967fa0, 1000, 0, NULL, NULL) = -1 EAGAIN (Resource
> temporarily unavailable)
>
> epoll_create1(EPOLL_CLOEXEC) = 3
>
> epoll_ctl(3, EPOLL_CTL_ADD, 11, {EPOLLIN|EPOLLERR|EPOLLHUP, {u32=31860176,
> u64=31860176}}) = 0
>
> epoll_ctl(3, EPOLL_CTL_ADD, 7, {EPOLLIN|EPOLLERR|EPOLLHUP, {u32=31860200,
> u64=31860200}}) = 0
>
> epoll_ctl(3, EPOLL_CTL_ADD, 10, {EPOLLIN|EPOLLERR|EPOLLHUP, {u32=31860224,
> u64=31860224}}) = 0
>
> epoll_wait(3, [{EPOLLIN, {u32=31860224, u64=31860224}}], 1, -1) = 1
>
> close(3) = 0
>
> recvfrom(10, "\2\0\0\0\250\3\0\0\7(at)\0\0\10\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
> 1000, 0, NULL, NULL) = 936
>
> recvfrom(10,
> "\2\0\0\0\250\3\0\0\0\0\0\0\10\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
> 1000, 0, NULL, NULL) = 936
>
> recvfrom(10, "\2\0\0\0x\1\0\0\7(at)\0\0\3\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
> 1000, 0, NULL, NULL) = 376
>
> recvfrom(10, 0x7ffeeb967fa0, 1000, 0, NULL, NULL) = -1 EAGAIN (Resource
> temporarily unavailable)
>
> epoll_create1(EPOLL_CLOEXEC) = 3
>
> epoll_ctl(3, EPOLL_CTL_ADD, 11, {EPOLLIN|EPOLLERR|EPOLLHUP, {u32=31860176,
> u64=31860176}}) = 0
>
> epoll_ctl(3, EPOLL_CTL_ADD, 7, {EPOLLIN|EPOLLERR|EPOLLHUP, {u32=31860200,
> u64=31860200}}) = 0
>
> epoll_ctl(3, EPOLL_CTL_ADD, 10, {EPOLLIN|EPOLLERR|EPOLLHUP, {u32=31860224,
> u64=31860224}}) = 0
>
> epoll_wait(3, [{EPOLLIN, {u32=31860224, u64=31860224}}], 1, -1) = 1
>
> close(3)
>
>
> Regards,
>
> Nikhil
>
> On Wed, Apr 27, 2022 at 8:25 PM Adelino Silva <adelino(dot)silva(at)pt(dot)ibm(dot)com>
> wrote:
>
> One possible cause for this problem is that the statistics collector
> process is bound to an IP:port which is not responding.
> See the following thread discussion.
>
>
> https://stackoverflow.com/questions/46008372/using-stale-statistics-instead-of-current-ones
>
> <https://stackoverflow.com/questions/46008372/using-stale-statistics-instead-of-current-ones>
> Using stale statistics instead of current ones - Stack Overflow
> <https://stackoverflow.com/questions/46008372/using-stale-statistics-instead-of-current-ones>
> Teams. Q&A for work. Connect and share knowledge within a single location
> that is structured and easy to search. Learn more
> stackoverflow.com
>
> Regards,
>
> Adelino Silva
>
> ------------------------------
> *From:* Nikhil Shetty <nikhil(dot)dba04(at)gmail(dot)com>
> *Sent:* Wednesday, April 27, 2022 2:49 PM
> *To:* Adelino Silva <adelino(dot)silva(at)pt(dot)ibm(dot)com>
> *Cc:* Pgsql-admin <pgsql-admin(at)lists(dot)postgresql(dot)org>
> *Subject:* [EXTERNAL] Re: Postgres Stale Statistics
>
> Hi Adelino, I had gone through that thread before, we cannot move the
> stats to RAM as of now. Thanks, Nikhil On Wed, Apr 27, 2022 at 6:16 PM
> Adelino Silva <adelino(dot)silva(at)pt(dot)ibm(dot)com> wrote: Hi, Found this thread
> that explains the warning.
> ZjQcmQRYFpfptBannerStart
> This Message Is From an External Sender
> This message came from outside your organization.
>
> ZjQcmQRYFpfptBannerEnd
> Hi Adelino,
>
> I had gone through that thread before, we cannot move the stats to RAM as
> of now.
>
> Thanks,
> Nikhil
>
> On Wed, Apr 27, 2022 at 6:16 PM Adelino Silva <adelino(dot)silva(at)pt(dot)ibm(dot)com>
> wrote:
>
> Hi,
>
> Found this thread that explains the warning.
> using stale statistics instead of current ones because stats collector is
> not responding
>
> https://www.postgresql.org/message-id/1457523467.24545.43.camel@2ndquadrant.com
>
> <https://www.postgresql.org/message-id/1457523467.24545.43.camel@2ndquadrant.com>
> PostgreSQL: Re: using stale statistics instead of current ones because
> stats collector is not responding
> <https://www.postgresql.org/message-id/1457523467.24545.43.camel@2ndquadrant.com>
> Hi, On Tue, 2016-03-08 at 16:18 -0800, Tory M Blue wrote: > No hits on the
> intratubes on this. > …
> www.postgresql.org
>
>
> Regards,
>
> Adelino Silva
>
> ------------------------------
> *From:* Nikhil Shetty <nikhil(dot)dba04(at)gmail(dot)com>
> *Sent:* Wednesday, April 27, 2022 12:08 PM
> *To:* Pgsql-admin <pgsql-admin(at)lists(dot)postgresql(dot)org>
> *Subject:* [EXTERNAL] Postgres Stale Statistics
>
> Hi, We are getting below WARNING on one of the standby instances. Not sure
> what caused it but to resolve it we tried restarting the database instances
> but it is still not working WARNING - using stale statistics instead of
> current ones because
> ZjQcmQRYFpfptBannerStart
> This Message Is From an External Sender
> This message came from outside your organization.
>
> ZjQcmQRYFpfptBannerEnd
> Hi,
>
> We are getting below WARNING on one of the standby instances. Not sure
> what caused it but to resolve it we tried restarting the database instances
> but it is still not working
>
>
> WARNING - using stale statistics instead of current ones because stats
> collector is not responding
>
>
> Postgresql version - 11.7
>
>
> Any other option to resolve this? We are thinking of building the standby
> again but what if the WARNING is for a primary database instance and a
> restart won't solve it?
>
>
> Thanks and Regards,
>
> Nikhil
>
>