On Tue, Aug 20, 2013 at 10:40 AM, Nourredine Rouibah < nourredine.roui...@gmail.com> wrote:
> Hi, > > After rebuilding Apache with the patch on port.c, is it correct to just > replace libapr-1.so.0.3.7 > in the running apache installation ? That would speed up the fix. > You have to * make a backup copy of the library * stop httpd * replace the library * start httpd again. Since it hangs intermittently anyway, pretend it is hung and has to be restarted ;) > > thanks, > > /Nourrédine. > > > On Tue, Aug 20, 2013 at 2:51 PM, Jeff Trawick <traw...@gmail.com> wrote: > >> On Tue, Aug 20, 2013 at 8:00 AM, Nourredine Rouibah < >> nourredine.roui...@gmail.com> wrote: >> >>> Hi, >>> >>> After doing some search I suspect our problems to be related with this >>> issue : https://issues.apache.org/bugzilla/show_bug.cgi?id=47645 >>> >>> I say that because the problem started to happen as soon as we changed >>> our Solaris HW to a new one. >>> Before that, we experienced http hang outs but once every 3 months and a >>> restart did solve the issue. >>> Since we moved to the new HW (same Solaris version) we httpd hangs many >>> times per day. >>> >>> As I described, the port 20000 hangs and a request to port 20001 did >>> release the httpd. >>> >>> Can someone confirm that it worth trying to rebuild with the above patch >>> ? >>> >> >> yes, it is worth getting that fix to try to resolve unresponsive behavior >> on Solaris with multiple listening sockets and a server that is not at max >> connections >> >> please just update the bloody software ;) that bug is ancient history >> >> (good work on the bug research; I had forgotten about that) >> >> > > > -- > Nourredine Rouibah > nourredine.roui...@gmail.com > -- Born in Roswell... married an alien... http://emptyhammock.com/