Jeffrey E Burgoyne wrote: > > > >> > >> Actually I think the behavior is the same, but the entry > >> in the logs > >> is different. I know in 1.3 it was chronological, which > >> leads me to > >> believe it was actually the finish time not the end time. > > > > Where is the difference between the finish time and the end time ? > > Isn't that the same ? > > Sorry, finish time not start time. you mean start time not finish time ? > > > > > I think i've been the victim of a DOS, realised with a SYN > > Flood. I'm > > not sure, but if yes, that's the first time. > > Currently i try to understand what happened. I have an entry in > > error_log which hit me on that idea: > > "[Sat Apr 23 15:19:02 2011] [error] server reached > > MaxClients setting, > > consider raising the MaxClients setting" > > Pretty sure a syn flood would not do that. does a typ syn on port 80 starts already a new child process ? > > I'll bet that for some reason the connections are being kept > open. Could be anything, even as simple as a bad keep alive session. > > Use the server-status page to see where the connections are > being held open. It has some good diagnostics, including a > column for how long the request has been served for. > > > > > > This is strange because this server has very rare visits, normally > > just my monitoring tool which checks the webserver every > > two minutes. > > This tool (ServersAlive, a windows application) never > > caused any problems. > > And beginning from saturday 15:30 the tool complains that it can't > > access the webserver. > > I had a lot of sockets on the webserver: > > > > tcp 133 0 146.107.x.x:80 146.107.135.80:3134 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3370 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3102 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3351 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3083 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3333 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3170 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3152 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3393 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3247 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3218 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3059 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3312 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3293 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3030 > CLOSE_WAIT > > - > > tcp 133 0 146.107.x.x:80 146.107.135.80:3269 > CLOSE_WAIT > > - > > ... And i also had 150 httpd processes running, + the one from root to open port 80. > > OK, based on this and what you wrote below there is something > with the networking, I am pretty sure. Basically you are > waiting on IO, which can be network, not just disk. Plus all > those open connections for what looks likes hours doing > nothing is also a dead give away. And yes that load average > is very high, just waiting for IO. > > > > > > > The IP 146.107.135.80 is the WinBox my monitoring tool is > running on. > > > > > > Also "top" on the linux box running the apache was very interesting: > > > > top - 13:17:45 up 25 days, 23:40, 4 users, load average: 155.00, > > 155.01, > > 155.03 > > Tasks: 587 total, 1 running, 586 sleeping, 0 stopped, 0 zombie > > Cpu(s): 0.2% us, 0.3% sy, 0.0% ni, 0.0% id, 99.5% wa, > 0.0% hi, > > 0.0% si > > Mem: 2075564k total, 1802740k used, 272824k free, > > 130936k buffers > > Swap: 1028152k total, 0k used, 1028152k free, > > 914700k cached > > > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > > 12820 root 17 0 2028 1236 1540 R 1.0 0.1 0:00.23 top > > 1 root 16 0 588 240 444 S 0.0 0.0 0:22.23 init > > 2 root RT 0 0 0 0 S 0.0 0.0 > > 0:00.00 migration/0 > > 3 root 34 19 0 0 0 S 0.0 0.0 > > 0:00.22 ksoftirqd/0 > > Bernd Helmholtz Zentrum MÃnchen Deutsches Forschungszentrum fÃr Gesundheit und Umwelt (GmbH) IngolstÃdter Landstr. 1 85764 Neuherberg www.helmholtz-muenchen.de Aufsichtsratsvorsitzende: MinDirÂin BÃrbel Brumme-Bothe GeschÃftsfÃhrer: Prof. Dr. GÃnther Wess und Dr. Nikolaus Blum Registergericht: Amtsgericht MÃnchen HRB 6466 USt-IdNr: DE 129521671
|