Hello, I keep errors "Bad file descriptor: apr_socket_opt_set: (TCP_NODELAY)" with apache wich freeze for few minutes...
I have a small charge on the server : 1.9 requests/sec. 8 Multiplexer and maxprocessors between 10 and 20 per virtualhost. I have to match the default user for apache server and user of Multiplexer, else error in starting apache. I use a specific user for default user for apache (no apache & no bobody), this settings can cause troubles or errors for peruser module ? Thank's Janno Sannik wrote: > What's you load (hits/day)? > We are running apache with abiut 1.2M hits/day and we sometimes need > 12-16 multiplexers to handle this (we use variable multiplexer patch, so > system adjusts multiplexers dynamically as needed). > > > On 20.03.2009 16:23, aguy wrote: >> Hi, >> >> I'm back with sad news, Ive same problems... my error log from apache : >> [Wed Mar 18 19:39:40 2009] [warn] (9)Bad file descriptor: >> apr_socket_opt_set: (TCP_NODELAY) >> [Wed Mar 18 22:04:41 2009] [warn] (9)Bad file descriptor: >> apr_socket_opt_set: (TCP_NODELAY) >> [Wed Mar 18 22:04:41 2009] [warn] (9)Bad file descriptor: >> apr_socket_opt_set: (TCP_NODELAY) >> [Thu Mar 19 07:48:04 2009] [warn] (9)Bad file descriptor: >> apr_socket_opt_set: (TCP_NODELAY) >> >> I'm increasing the number of multiplexer counts for 6 to see ? >> >> Anybody can tell me what links they are between : >> - multiplexer >> - MinSpareProcessors >> - MinProcessors >> - MaxProcessors >> - ServerLimit (MaxClients is the same ?) >> >> ?? thank's a lot !! >> >> >> aguy wrote: >> >>> Hi, >>> >>> Apache frozen for few minutes during this warm, I'm increasing the number >>> of multiplexer counts for 4. >>> I'm waiting and see for this point... >>> Normaly each day We have 1 freeze during 5~10 mns, I will return the result >>> of this modification. >>> >>> Thanks's for your answer ! >>> >>> >>> Taavi Sannik wrote: >>> >>>> Hi! >>>> >>>>> [Wed Mar 18 22:04:41 2009] [warn] (9)Bad file descriptor: >>>>> apr_socket_opt_set: (TCP_NODELAY) >>>>> [Thu Mar 19 07:48:04 2009] [warn] (9)Bad file descriptor: >>>>> apr_socket_opt_set: (TCP_NODELAY) >>>>> >>>> We get this too, but we don't experience any problems (at least for now). >>>> >>>> >>>>> and I don't understand why in my server-status I have some process >>>>> describes with : >>>>> 20 0 STANDBY UNKNOWN -1 -1 (null) >>>>> >>>>> Why process runs under uid -1 and gid -1 ?? >>>>> >>>> This actually means the child is not running anymore, so it's completely >>>> normal. >>>> >>>> Have you tried increasing the multiplexer count ? >>>> >>>> >>>> >>>> -- >>>> Taavi >>>> >>>> >>>> ------------------------------------------------------------------------ >>>> >>>> _______________________________________________ >>>> Peruser mailing list >>>> [email protected] >>>> http://www.telana.com/mailman/listinfo/peruser >>>> >> >> _______________________________________________ >> Peruser mailing list >> [email protected] >> http://www.telana.com/mailman/listinfo/peruser >> > > > ------------------------------------------------------------------------ > > _______________________________________________ > Peruser mailing list > [email protected] > http://www.telana.com/mailman/listinfo/peruser _______________________________________________ Peruser mailing list [email protected] http://www.telana.com/mailman/listinfo/peruser
