From owner-freebsd-stable Fri Sep 21 15: 8:11 2001 Delivered-To: freebsd-stable@freebsd.org Received: from acura.isprime.com (acura.isprime.com [130.94.138.66]) by hub.freebsd.org (Postfix) with ESMTP id 5ABE937B403 for ; Fri, 21 Sep 2001 15:07:50 -0700 (PDT) Received: from energy (localhost [127.0.0.1]) by acura.isprime.com (8.11.3/8.11.2) with SMTP id f8LM7ml79917; Fri, 21 Sep 2001 18:07:48 -0400 (EDT) Message-ID: <001f01c142e9$cc0f9130$02010a0a@energy> From: "ISPrime Support" To: "Luke Cowell" , References: Subject: Re: memory leak in FBSD 4.4-STABLE Date: Fri, 21 Sep 2001 18:07:25 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_001C_01C142C8.44A549E0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2600.0000 Disposition-Notification-To: "ISPrime Support" X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG This is a multi-part message in MIME format. ------=_NextPart_000_001C_01C142C8.44A549E0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Re: memory leak in FBSD 4.4-STABLEI already did that on the boxes and = the configs are identical. --Phil Rosenthal Team Supreme, Inc. 347-739-4737 ----- Original Message -----=20 From: Luke Cowell=20 To: ISPrime Support ; stable@FreeBSD.ORG=20 Sent: Friday, September 21, 2001 6:04 PM Subject: Re: memory leak in FBSD 4.4-STABLE Well at any rate here's something you can do to keep you box stable = until a real solution in found. Update the MaxRequestsPerChild directive = in the apache conf. This may be at 0 where a child process will never = die. Set it to something like 100 so the child will die after it serves = 100 requests. I would go through the config and compare everything in the first = section on the 2 different servers.=20 Luke=20 On 9/21/01 2:47 PM, "ISPrime Support" wrote: Same binary of apache on both boxes, 1.3.20 --Phil Rosenthal Team Supreme, Inc. 347-739-4737 ----- Original Message -----=20 From: Luke Cowell =20 To: ISPrime Support =20 Sent: Friday, September 21, 2001 5:44 PM Subject: Re: memory leak in FBSD 4.4-STABLE On 9/21/01 2:35 PM, "ISPrime Support" wrote: I have two servers load balanced running a large-load website = (about 20 megabit each server) one is running 4.4-STABLE other is 4.3-STABLE from about two = months ago. the apache binary is the same on both. the 4.4 system over time the apache processes get HUGE until the = system runs out of swap, and the 4.3 system stays small. last pid: 55332; load averages: 64.88, 25.57, 17.49 up = 48+04:46:20 17:29:09 768 processes: 13 running, 754 sleeping, 1 zombie CPU states: 24.0% user, 0.0% nice, 18.4% system, 8.6% = interrupt, 48.9% idle Mem: 574M Active, 91M Inact, 317M Wired, 20M Cache, 112M Buf, = 1664K Free Swap: 3072M Total, 128M Used, 2944M Free, 4% Inuse, 16K In, 13M = Out =20 PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU = CPU COMMAND 41660 nobody 2 0 6996K 4452K select 1 0:03 10.25% = 10.25% httpd-statu 41543 nobody 2 0 6688K 4192K select 0 0:03 9.77% = 9.77% httpd-statu 42025 nobody 2 0 6400K 3760K accept 1 0:03 9.72% = 8.84% httpd-statu 55332 root 24 -20 2800K 880K CPU1 1 0:02 22.13% = 6.54% top 41792 nobody 2 0 7056K 4492K select 1 0:03 4.44% = 4.44% httpd-statu 41872 nobody 2 0 5980K 3804K select 1 0:02 4.35% = 4.35% httpd-statu 3 root 18 0 0K 0K psleep 0 38:01 4.30% = 4.30% vmdaemon 41553 nobody 2 0 5980K 3428K accept 0 0:03 4.72% = 4.30% httpd-statu 41915 nobody 2 0 7648K 4912K accept 0 0:04 4.20% = 4.20% httpd-statu 42040 nobody 2 0 5276K 3336K select 1 0:02 4.15% = 4.15% httpd-statu 41736 nobody 2 0 6112K 3744K select 1 0:05 4.10% = 4.10% httpd-statu 42007 nobody 2 0 5980K 3508K select 1 0:04 4.46% = 4.05% httpd-statu 41847 nobody 2 0 4700K 2492K accept 0 0:02 2.68% = 2.44% httpd-statu This is after about 20 minutes of running. given 3-4 hours they = will be about 10 megs each. This is the 4.3 system after 1 hour of uptime: last pid: 23347; load averages: 3.58, 1.88, 1.72 up = 3+16:15:28 17:35:16 1328 processes:12 running, 1315 sleeping, 1 zombie CPU states: 11.7% user, 0.0% nice, 8.1% system, 6.4% = interrupt, 73.8% idle Mem: 661M Active, 55M Inact, 237M Wired, 50M Cache, 112M Buf, = 1664K Free Swap: 1024M Total, 8960K Used, 1015M Free PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU = COMMAND 182 mysql 2 0 12352K 156K poll 0 0:05 0.00% = 0.00% mysqld 23344 root -10 -20 3440K 940K CPU0 0 0:03 1.37% = 1.37% top 195 nobody 2 0 3308K 248K accept 0 0:00 0.00% = 0.00% phphttpd 196 nobody 2 0 3308K 248K accept 0 0:00 0.00% = 0.00% phphttpd 193 nobody 2 0 3304K 236K accept 0 0:00 0.00% = 0.00% phphttpd 194 nobody 2 0 3296K 232K accept 0 0:00 0.00% = 0.00% phphttpd 191 root 2 0 3056K 208K select 0 0:18 0.00% = 0.00% phphttpd 23337 root 2 0 2308K 1324K select 1 0:00 0.00% = 0.00% sshd 19526 root 2 0 2308K 1276K select 1 0:00 0.00% = 0.00% sshd 106 root 2 0 2224K 528K select 0 0:01 0.00% = 0.00% sshd 8049 nobody 2 0 2068K 1236K sbwait 0 0:01 0.00% = 0.00% httpd-status 7622 nobody 2 0 2064K 1268K select 0 0:01 0.00% = 0.00% httpd-status 8251 nobody 2 0 2056K 1224K sbwait 0 0:01 0.00% = 0.00% httpd-status 7260 nobody 2 0 2056K 1224K sbwait 0 0:01 0.00% = 0.00% httpd-status 7780 nobody 2 0 2056K 1224K select 1 0:01 0.00% = 0.00% httpd-status 7898 nobody 2 0 2056K 1204K select 0 0:01 0.00% = 0.00% httpd-status =20 Any ideas? Yeah, I'd check what versions of apache are you running ? You = theory doesn't hold much water unless the version are the same. I would = also check to see what sort of memory configuration you have for apache. Luke=20 ------=_NextPart_000_001C_01C142C8.44A549E0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Re: memory leak in FBSD 4.4-STABLE
I already did that on the boxes and the = configs are=20 identical.
--Phil Rosenthal
Team Supreme, Inc.
347-739-4737
----- Original Message -----
From:=20 Luke Cowell
To: ISPrime Support ; stable@FreeBSD.ORG
Sent: Friday, September 21, = 2001 6:04=20 PM
Subject: Re: memory leak in = FBSD=20 4.4-STABLE

Well at any rate here=92s something = you can do=20 to keep you box stable until a real solution in found. Update the=20 MaxRequestsPerChild directive in the apache conf. This may be at 0 = where a=20 child process will never die. Set it to something like 100 so the = child will=20 die after it serves 100 requests.

 I would go through the = config=20 and compare everything in the first section on the 2 different = servers.=20

Luke


On 9/21/01 2:47 = PM,=20 "ISPrime Support" <support@isprime.com> wrote:
Same binary of apache on = both=20 boxes, 1.3.20
--Phil = Rosenthal
Team=20 Supreme, Inc.
347-739-4737
----- Original Message ----- =
From: Luke Cowell=20 <mailto:lukec@georgio.e-and-s.com>  
To: = ISPrime=20 Support <mailto:support@isprime.com>  
Sent: = Friday,=20 September 21, 2001 5:44 PM
Subject: Re: memory leak in = FBSD=20 4.4-STABLE

On 9/21/01 2:35 PM, "ISPrime Support"=20 <support@isprime.com> wrote:
I have two servers = load=20 balanced running a large-load website (about 20 megabit each=20 server)
one is running 4.4-STABLE other is 4.3-STABLE from = about two=20 months ago.
the apache binary is the same on both.
the 4.4 = system=20 over time the apache processes get HUGE until the system runs = out of=20 swap, and the 4.3 system stays small.
last pid: 55332; =  load=20 averages: 64.88, 25.57, 17.49  up 48+04:46:20 =  17:29:09
768=20 processes: 13 running, 754 sleeping, 1 zombie
CPU states: = 24.0% user,=20  0.0% nice, 18.4% system,  8.6% interrupt, 48.9% = idle
Mem:=20 574M Active, 91M Inact, 317M Wired, 20M Cache, 112M Buf, 1664K=20 Free
Swap: 3072M Total, 128M Used, 2944M Free, 4% Inuse, 16K = In, 13M=20 Out
  
  PID USERNAME PRI NICE =  SIZE=20    RES STATE  C   TIME =   WCPU=20    CPU COMMAND
41660 nobody =     2=20   0  6996K  4452K select 1   0:03 = 10.25%=20 10.25% httpd-statu
41543 nobody     2=20   0  6688K  4192K select 0   0:03=20  9.77%  9.77% httpd-statu
42025 nobody=20     2   0  6400K  3760K = accept 1=20   0:03  9.72%  8.84% httpd-statu
55332 = root=20      24 -20  2800K =   880K CPU1=20   1   0:02 22.13%  6.54% top
41792 = nobody=20     2   0  7056K  4492K = select 1=20   0:03  4.44%  4.44% httpd-statu
41872 = nobody=20     2   0  5980K  3804K = select 1=20   0:02  4.35%  4.35%=20 httpd-statu
    3 root=20      18   0=20     0K     0K psleep 0=20  38:01  4.30%  4.30% vmdaemon
41553 nobody=20     2   0  5980K  3428K = accept 0=20   0:03  4.72%  4.30% httpd-statu
41915 = nobody=20     2   0  7648K  4912K = accept 0=20   0:04  4.20%  4.20% httpd-statu
42040 = nobody=20     2   0  5276K  3336K = select 1=20   0:02  4.15%  4.15% httpd-statu
41736 = nobody=20     2   0  6112K  3744K = select 1=20   0:05  4.10%  4.10% httpd-statu
42007 = nobody=20     2   0  5980K  3508K = select 1=20   0:04  4.46%  4.05% httpd-statu
41847 = nobody=20     2   0  4700K  2492K = accept 0=20   0:02  2.68%  2.44%=20 httpd-statu

This is after about 20 minutes of = running.=20  given 3-4 hours they will be about 10 megs=20 each.

This is the 4.3 system after 1 hour of = uptime:
last pid:=20 23347;  load averages:  3.58,  1.88,  1.72=20    up 3+16:15:28  17:35:16
1328 = processes:12=20 running, 1315 sleeping, 1 zombie
CPU states: 11.7% user, =  0.0%=20 nice,  8.1% system,  6.4% interrupt, 73.8% = idle
Mem: 661M=20 Active, 55M Inact, 237M Wired, 50M Cache, 112M Buf, 1664K = Free
Swap:=20 1024M Total, 8960K Used, 1015M Free

PID = USERNAME PRI=20 NICE  SIZE    RES STATE  C =   TIME=20   WCPU    CPU = COMMAND
  182 mysql=20      2   0 12352K =   156K=20 poll   0   0:05  0.00%  0.00%=20 mysqld
23344 root     -10 -20  3440K =   940K CPU0   0   0:03  1.37% =  1.37% top
  195 nobody =     2=20   0  3308K   248K accept 0 =   0:00=20  0.00%  0.00% phphttpd
  196 nobody=20     2   0  3308K =   248K=20 accept 0   0:00  0.00%  0.00%=20 phphttpd
  193 nobody     2=20   0  3304K   236K accept 0 =   0:00=20  0.00%  0.00% phphttpd
  194 nobody=20     2   0  3296K =   232K=20 accept 0   0:00  0.00%  0.00%=20 phphttpd
  191 root =       2=20   0  3056K   208K select 0 =   0:18=20  0.00%  0.00% phphttpd
23337 root=20       2   0  2308K=20  1324K select 1   0:00  0.00%  0.00%=20 sshd
19526 root       2 =   0=20  2308K  1276K select 1   0:00  0.00%=20  0.00% sshd
  106 root=20       2   0  2224K=20   528K select 0   0:01  0.00% =  0.00%=20 sshd
 8049 nobody     2 =   0=20  2068K  1236K sbwait 0   0:01  0.00%=20  0.00% httpd-status
 7622 nobody =     2=20   0  2064K  1268K select 0   0:01=20  0.00%  0.00% httpd-status
 8251 nobody=20     2   0  2056K  1224K = sbwait 0=20   0:01  0.00%  0.00% = httpd-status
 7260=20 nobody     2   0  2056K =  1224K=20 sbwait 0   0:01  0.00%  0.00%=20 httpd-status
 7780 nobody     2=20   0  2056K  1224K select 1   0:01=20  0.00%  0.00% httpd-status
 7898 nobody=20     2   0  2056K  1204K = select 0=20   0:01  0.00%  0.00%=20 httpd-status

 
Any=20 ideas?



Yeah, I=92d=20 check what versions of apache are you running ? You theory = doesn=92t hold=20 much water unless the version are the same. I would also check to = see what=20 sort of memory configuration you have for apache.

Luke=20


------=_NextPart_000_001C_01C142C8.44A549E0-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message