On Thu, Feb 2, 2012 at 10:08 AM, Markus Neteler <neteler@osgeo.org> wrote:
...
Fine, the GRASS Wiki is back, too...
Remaining problem:
/var/log/apache2/error.log
...
/osgeo/grass/grass-wiki/bin/ulimit4.sh: xmalloc: ../bash/subst.c:3503:
cannot allocate 731 bytes (32768 bytes allocated)
What does that mean?
Playground for tests: http://grass.osgeo.org/wiki/Sandbox
thanks for help,
Markus
On 02/14/2012 10:54 AM, Markus Neteler wrote:
On Thu, Feb 2, 2012 at 10:08 AM, Markus Neteler <neteler@osgeo.org> wrote:
...Fine, the GRASS Wiki is back, too...
Remaining problem:
/var/log/apache2/error.log
...
/osgeo/grass/grass-wiki/bin/ulimit4.sh: xmalloc: ../bash/subst.c:3503:
cannot allocate 731 bytes (32768 bytes allocated)What does that mean?
Playground for tests: http://grass.osgeo.org/wiki/Sandbox
thanks for help,
Markus
I bumped gunicorn down from 8 to 4 workers to see if that helps.
http://wiki.osgeo.org/wiki/Spatialreference.org
While still above pre upgrade stats the current readouts look much
better and I haven't seen any user complaints on IRC
http://webextra.osgeo.osuosl.org/munin/osgeo.org/projects.osgeo.org.html
Thanks,
Alex
Alex wrote:
While still above pre upgrade stats the current readouts look much
better and I haven't seen any user complaints on IRC
http://webextra.osgeo.osuosl.org/munin/osgeo.org/projects.osgeo.org.html
something still ain't right.
ksoftirqd remains pegged near 100%, making the system less than responsive.
a flood of incoming network connections could explain this, or some VM/host
feedback issue, or...?
hmmm, dmesg is full of:
TCPv6: Possible SYN flooding on port 80. Dropping request
munin's monitoring has been dropping out as well:
http://webextra.osgeo.osuosl.org/munin/osgeo.org/projects.osgeo.org.html
Some iptables voodoo seems in order.
regards,
Hamish
One might think that the iptables blacklist is probably growing too
large. We might drop all entries and start from scratch by just
blocking those who still didn't notice.
Done,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
Martin wrote:
One might think that the iptables
blacklist is probably growing too
large. We might drop all entries and start from
scratch by just blocking those who still didn't notice.
the terminal session was getting really slow there for a little while,
but now all of a sudden everything has gone back to normal, memory use
is back down & no more swapping, ksoftirqd and events no longer pegging
the cpu, load has returned to near 0.1 ..
?!,
Hamish
On Thu, Feb 16, 2012 at 8:46 PM, Martin Spott <Martin.Spott@mgras.net> wrote:
One might think that the iptables blacklist is probably growing too
large. We might drop all entries and start from scratch by just
blocking those who still didn't notice.
Sounds good!
Markus