Thread: Shared memory changes in 9.4?
Hi,
I've been trying to take the 9.4 beta for a spin, but seem to have run into a shared memory issue on startup:
FATAL: could not open shared memory segment "/PostgreSQL.1804289383": Permission denied
We're running Postgres inside LXC, which I suspect is causing problems here, but 9.0 through 9.3 run just fine in the same configuration. Any idea what may have changed? I suspect the solution here is tweaking some LXC setting, but I was hoping someone can point me in the right direction.
Thanks,
Maciek Sakrejda
Heroku Postgres
Maciek Sakrejda <m.sakrejda@gmail.com> writes: > I've been trying to take the 9.4 beta for a spin, but seem to have run into > a shared memory issue on startup: > FATAL: could not open shared memory segment "/PostgreSQL.1804289383": > Permission denied > We're running Postgres inside LXC, which I suspect is causing problems > here, but 9.0 through 9.3 run just fine in the same configuration. Any idea > what may have changed? I suspect the solution here is tweaking some LXC > setting, but I was hoping someone can point me in the right direction. This message is coming out of the new "dynamic shared memory" code. I'm not real sure why that's being invoked at startup; perhaps Robert Haas can elucidate. Anyway the short answer is that this sounds like a portability hazard in the new code in src/backend/storage/ipc/dsm_impl.c. Perhaps you can look into that and identify what's up. regards, tom lane
Hi, On 2014-05-25 12:26:20 -0700, Maciek Sakrejda wrote: > I've been trying to take the 9.4 beta for a spin, but seem to have run into > a shared memory issue on startup: > > FATAL: could not open shared memory segment "/PostgreSQL.1804289383": > Permission denied > > We're running Postgres inside LXC, which I suspect is causing problems > here, but 9.0 through 9.3 run just fine in the same configuration. Any chance you're using a 9.3 configuration file instead of the one generated by initdb? dynamic_shared_memory_type defaults to 'posix' if not specified in the config file (on platforms supporting it). If initdb detects that 'posix' can't be used it'll emit a different value. If you're copying the config from 9.3 and your environment doesn't support posix shm that'll cause the above error. I still think dynamic_shared_memory_type should default to 'none' because of such problems. > Any idea > what may have changed? I suspect the solution here is tweaking some LXC > setting, but I was hoping someone can point me in the right direction. It sounds like you're not allowing posix shared memory. Possibly just by not mounting /dev/shm. Greetings, Andres Freund -- Andres Freund http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Training & Services
On Tue, May 27, 2014 at 8:22 PM, Maciek Sakrejda <m.sakrejda@gmail.com> wrote: > On Mon, May 26, 2014 at 12:24 AM, Andres Freund <andres@2ndquadrant.com> > wrote: >> >> Any chance you're using a 9.3 configuration file instead of the one >> generated by initdb? >> dynamic_shared_memory_type defaults to 'posix' if not specified in the >> config file (on platforms supporting it). If initdb detects that 'posix' >> can't be used it'll emit a different value. If you're copying the config >> from 9.3 and your environment doesn't support posix shm that'll cause >> the above error. >> I still think dynamic_shared_memory_type should default to 'none' >> because of such problems > > It works with 'none' and 'sysv'--I think the issue is that technically our > environment does support 'posix', but '/dev/shm' is indeed not mounted in > the LXC container, leading to a discrepancy between what initdb decides and > what's actually possible. Thanks for your help. I think it would be good to understand why initdb isn't getting this right. Did you run initdb outside the LXC container, where /dev/shm would have worked, but then run postgres inside the LXC container, where /dev/shm does not work? I ask because initdb is supposed to be doing the same thing that postgres does, so it really ought to come to the same conclusion about what will and won't work. With regard to Andres' proposal, I'm not that keen on setting dynamic_shared_memory_type='none' by default. Would we leave it that way until we get in-core users of the facility, and then change it? I guess that'd be OK, but frankly if enabling dynamic_shared_memory_type by default is causing us many problems, then we'd better reconsider the design of the facility now, before we start adding more dependencies on it. We've already fixed a bunch of DSM-related issues as a result of the fact that the default *isn't* none, and I dunno how many of those we would have found if the default had been none. I tend to think DSM is an important facility that we're going to be wanting to build on in future releases, so I'm keen to have it available by default so that we can iron out any kinks before we get too far down that path. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
On Tue, May 27, 2014 at 8:23 PM, Robert Haas <robertmhaas@gmail.com> wrote:
>
> I think it would be good to understand why initdb isn't getting this
> right. Did you run initdb outside the LXC container, where /dev/shm
> would have worked, but then run postgres inside the LXC container,
> where /dev/shm does not work? I ask because initdb is supposed to be
> doing the same thing that postgres does, so it really ought to come to
> the same conclusion about what will and won't work.
You're absolutely right--I thought initdb was containerized as well, but
>
> I think it would be good to understand why initdb isn't getting this
> right. Did you run initdb outside the LXC container, where /dev/shm
> would have worked, but then run postgres inside the LXC container,
> where /dev/shm does not work? I ask because initdb is supposed to be
> doing the same thing that postgres does, so it really ought to come to
> the same conclusion about what will and won't work.
You're absolutely right--I thought initdb was containerized as well, but
I looked at our code and this is exactly what's happening.
> ....We've already fixed a bunch of DSM-related issues
> as a result of the fact that the default *isn't* none, and I dunno how
> many of those we would have found if the default had been none.
For what it's worth, +1. I'm not sure whether or not we had a good reason for
> ....We've already fixed a bunch of DSM-related issues
> as a result of the fact that the default *isn't* none, and I dunno how
> many of those we would have found if the default had been none.
For what it's worth, +1. I'm not sure whether or not we had a good reason for
doing initdb outside the container, but it's definitely an aberrant configuration,
and should not be taken as evidence that the current default is a problem.
On Mon, May 26, 2014 at 12:24 AM, Andres Freund <andres@2ndquadrant.com> wrote:
Any chance you're using a 9.3 configuration file instead of the onegenerated by initdb?
dynamic_shared_memory_type defaults to 'posix' if not specified in the
config file (on platforms supporting it). If initdb detects that 'posix'
can't be used it'll emit a different value. If you're copying the config
from 9.3 and your environment doesn't support posix shm that'll cause
the above error.
I still think dynamic_shared_memory_type should default to 'none'
because of such problems.
It works with 'none' and 'sysv'--I think the issue is that technically our environment does support 'posix', but '/dev/shm' is indeed not mounted in the LXC container, leading to a discrepancy between what initdb decides and what's actually possible. Thanks for your help.
Thanks,
Maciek