Hi,

currently i m running apache(webserver)+mysql(databases)+java(application server) on 1 linux machine,which is server load heavy.

i would like to get another 2 dell poweredge and a 3com switch, and link up these 3 machine to do the 'balancing' and backup routine job for the databases.

Machine 1 running the webserver+application,
machine 2 just running the database,
machinse 3 doing the backup job (which is daily copying the database from machine 2 and some file from machine 1)

these 3 machine is connected to the switch.

The data flow will be like this:
web user request --> router --> switch
-> machine 1 (application open new connection to db)
-> goto switch
-> machine 2 (database take the query,process, return back)
->switch -> machine 1 -> switch -> router -> web user

is this the correct way ? or any better design and idea?

No, role of the switch is not clear. I can help you, if you give me the following information:
- host parameters (CPU, mem, disk(s))
- output from "top"
- complete line as you start application server
- release of JVM, Apache, Application Server(indicate which one Tomcat, JBoss, JRun), MySQL
- type of connector between Apache and Application Server
- connector between MySQL and Application Server

I may need (not now) output from logs.

Peter
http://opensolution-us.com

HI,

What i wanna to implement is :
linux machine 1 (192.168.1.1) dedicated for webserver and application
another linux machine 2 (192.168.1.2) dedicated to Mysql databases;
linux machine 3 (192.168.1.3) dedicated to copying the file and databases from linux 1 and linux 2 for backup.

these 3 linux machine is connected to a switch and connected to router.

will this the right way ?

That is just fine.

Backups done on the private lan, communication between web/app server and mysql server also on the private lan (but not on the same network interfaces used for the backups).

Go for it.

Originally posted by qazakax
HI,

What i wanna to implement is :
linux machine 1 (192.168.1.1) dedicated for webserver and application
another linux machine 2 (192.168.1.2) dedicated to Mysql databases;
linux machine 3 (192.168.1.3) dedicated to copying the file and databases from linux 1 and linux 2 for backup.

these 3 linux machine is connected to a switch and connected to router.

will this the right way ?

OK guys, let me work on this one.
It's nice you gave me ip addresses, but I need info about hardware/software configuration/resources to give you correct advice. Your problem is a perceived performance problem. Am I correct?
I need also answers on two more questions:
1. Do you have DNS server?
2. Do you have an e-mail server?

Peter
http://opensolution-us.com

hi,

the hardware for the 3 machine will be Dell poweredge,
2.4xeon, 2GBRAM,2XSCSI hardrive,

192.168.1.1 willbe running apache2, modperl, perl
192.168.1.2 willbe only running mysql, for data query and update
192.168.1.3 will doing the backup

DNS server will be using ISP DNS server, there is no emailserver.

Main reason is the 192.168.1.1, because currently our apache2+modperl setup take alot of memory, maybe because of modperl memory leakage ?..

Our current setup is Xeon2.4, 1GB ram, 1xSCSI,1XIDE, the webserver+modperl+mysql, all in a same server.

server load is ok, peak only up to 4.0, but the memory is used up alot.

03:58:03 up 24 days, 17:40, 1 user, load average: 0.07, 0.36, 0.35
88 processes: 85 sleeping, 3 running, 0 zombie, 0 stopped
CPU states: 0.0% user 0.2% system 0.0% nice 0.0% iowait 99.8% idle
Mem: 1030892k av, 879836k used, 151056k free, 0k shrd, 51976k buff
500808k actv, 9528k in_d, 304116k in_c
Swap: 1534168k av, 142508k used, 1391660k free 531496k cached

Did you just restart apache in that top capture? Otherwise, there's absolutely no reason the box should have 150MB of memory idle. mod_perl likes to allocate, but not deallocate, so it grows over time, even if there's no actual leak. You should look into lingerd or squid to reduce the number of apaches you need, which will ease your memory needs.

Adding more memory should help nicely. You can go with multiple boxes, but I'm not sure that's necessary.

Let us start with one machine easily identified. It's no brainier to allocate one machine solely for the DB use. As a matter of fact the DB machine typically is I/O bound with relatively low requirements for CPU and/or memory.

So, possible bottlenecks for good throughput must be identified and they are:
- disk I/O
- network/adapter I/O

I understand that each machine will have 2 hard disks. The typical SCSI-2 adapter can handle theoretically up to 20Mb/sec with disk about 5Mb/sec. So, even if you configure your disks as RAID-0 (maximum parallelization) your combined throughput of both disks ~10Mb/sec will be less than adapter. You must understand the concept of RAID-0,1,4,5 to make a right choice. I need more info about your application to give a correct advice.

Now, the network Ethernet adapter 10/100 ( I assume you have it) must be configured as full-duplex in 32-bit autonegotiation or directly with 100Mb/sec. This way it will not constitute a possible bottleneck. But the switch could be, pay very close attention to the type of switch you are buying.

This machine also must have a tape or CD-RW system for backup. There is no needs to dedicate one machine for this role. The second benefits with this approach is an ability to redirect traffic from the network to local SCSI bus.

Now, I think we need to finish with this machine before proceed to another machines configuration.

Any questions?



Peter
http://opensolution-us.com

Why would anyone use SCSI2 when ultra3 SCSI is available? Anyway I would run vmstat during peak hours to see exactly where your machine is lagging at. I would try just adding another gig of ram to your current server. That is a much cheaper solution then buying more servers which I think right now is overkill. When you posted those stats from top, was that during peak usage?

Originally posted by genlee
Why would anyone use SCSI2 when ultra3 SCSI is available? Anyway I would run vmstat during peak hours to see exactly where your machine is lagging at. I would try just adding another gig of ram to your current server. That is a much cheaper solution then buying more servers which I think right now is overkill. When you posted those stats from top, was that during peak usage?

Indeed, I am probably off with the SCSI-2 as well as with estimation of contemporary disk's throughput, but I didn't have complete input. And I was guessing. Now, I would ask you to give me a chance to finish what I started. Believe me, it will help few people including yourself.

The vmstat command in this context is not an exact option. In addition to vmstat command (which must be used with proper arguments) there are tons of another options, combination of which could produce (I emphases - "could") a desirable outcome. I can name few - sar, iostat, mpstat (all together sysstat in some implementations). The combination of these plus few more might give some system impression. This is much more recommended approach than "which I think right now is overkill". How you suppose to know without even your recommended vmstat output?

Any more questions?

Peter
http://opensolution-us.com

I really don't see how all those other utilities are so much better then vmstat. By watching vmstat during peak usage, I could tell within minutes what is causing the load on the server. You can use those other utilities if you want but they will just give you the same info that vmstat does. My approach was using vmstat to monitor the system during peak usage and adding more memory, not an out of context sentence fragment you posted. The server load is only getting up to 4, adding more memory and doing some apache tuning will fix that. And yes I do think that adding more servers to resolve a system load that low is overkill unless you have an unlimited budget.

Originally posted by genlee
I really don't see how all those other utilities are so much better then vmstat. By watching vmstat during peak usage, I could tell within minutes what is causing the load on the server. You can use those other utilities if you want but they will just give you the same info that vmstat does. My approach was using vmstat to monitor the system during peak usage and adding more memory, not an out of context sentence fragment you posted. The server load is only getting up to 4, adding more memory and doing some apache tuning will fix that. And yes I do think that adding more servers to resolve a system load that low is overkill unless you have an unlimited budget.

We will discuss this later.
Now, any more questions?

Peter

<<< Signature to be setup in your profile >>>

This is my server spec + apache2, modperl config + mysql config

What i m facing right now is, the apache2_modperl is not realiable, this will took up all the memory, system freeze.

Any way to scale n tune the performance?



#######################
ionnex kernel: Out of Memory: Killed process 4001 (httpd).
ionnex syslogd 1.4.1: restart.
ionnex syslog: syslogd startup succeeded
ionnex kernel: klogd 1.4.1, log source = /proc/kmsg started.
ionnex kernel: Linux version 2.4.20-28.9 (bhcompile@porky.devel.redhat.com) (gcc version 3.2.2 20030222 (Red Hat
Linux 3.2.2-5)) #1 Thu Dec 18 13:45:22 EST 2003
ionnex kernel: BIOS-provided physical RAM map:
ionnex kernel: BIOS-e820: 0000000000000000 - 000000000009f400 (usable)
ionnex kernel: BIOS-e820: 000000000009f400 - 00000000000a0000 (reserved)
ionnex kernel: BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved)
ionnex kernel: BIOS-e820: 0000000000100000 - 000000005f7f0000 (usable)
ionnex kernel: BIOS-e820: 000000005f7f0000 - 000000005f7f3000 (ACPI NVS)
ionnex kernel: BIOS-e820: 000000005f7f3000 - 000000005f800000 (ACPI data)
ionnex kernel: BIOS-e820: 00000000fec00000 - 0000000100000000 (reserved)
ionnex kernel: 631MB HIGHMEM available.
ionnex kernel: 896MB LOWMEM available.
ionnex syslog: klogd startup succeeded
ionnex kernel: On node 0 totalpages: 391152
ionnex kernel: zone(0): 4096 pages.
ionnex kernel: zone(1): 225280 pages.
ionnex kernel: zone(2): 161776 pages.
ionnex kernel: Kernel command line: ro root=LABEL=/
ionnex kernel: Initializing CPU#0
ionnex kernel: Detected 2798.702 MHz processor.
ionnex rpc.statd[3333]: Version 1.0.1 Starting
ionnex nfslock: rpc.statd startup succeeded
ionnex kernel: Console: colour VGA+ 80x25
ionnex kernel: Calibrating delay loop... 5583.66 BogoMIPS
ionnex kernel: Memory: 1535276k/1564608k available (1359k kernel code, 24720k reserved, 1004k data, 132k init, 6
47104k highmem)
ionnex kernel: Dentry cache hash table entries: 262144 (order: 9, 2097152 bytes)
ionnex kernel: Inode cache hash table entries: 131072 (order: 8, 1048576 bytes)
ionnex kernel: Mount cache hash table entries: 512 (order: 0, 4096 bytes)
ionnex kernel: Buffer-cache hash table entries: 131072 (order: 7, 524288 bytes)
ionnex keytable: Loading keymap:
ionnex kernel: Page-cache hash table entries: 524288 (order: 9, 2097152 bytes)
ionnex keytable:
ionnex kernel: CPU: Trace cache: 12K uops, L1 D cache: 8K
ionnex keytable:
ionnex kernel: CPU: L2 cache: 512K
ionnex rpc.statd[3333]: unable to register (statd, 1, udp).
ionnex rc: Starting keytable: succeeded
ionnex kernel: Intel machine check architecture supported.
ionnex kernel: Intel machine check reporting enabled on CPU#0.
ionnex random: Initializing random number generator: succeeded
ionnex kernel: CPU: Intel(R) Pentium(R) 4 CPU 2.80GHz stepping 07
ionnex kernel: Enabling fast FPU save and restore... done.
ionnex kernel: Enabling unmasked SIMD FPU exception support... done.
ionnex kernel: Checking 'hlt' instruction... OK.
ionnex rc: Starting pcmcia: succeeded
ionnex kernel: POSIX conformance testing by UNIFIX
ionnex kernel: mtrr: v1.40 (20010327) Richard Gooch (rgooch@atnf.csiro.au)
ionnex kernel: mtrr: detected mtrr type: Intel
ionnex kernel: PCI: PCI BIOS revision 2.10 entry at 0xfb310, last bus=1
ionnex kernel: PCI: Using configuration type 1
ionnex kernel: PCI: Probing PCI hardware
ionnex kernel: PCI: Ignoring BAR0-3 of IDE controller 00:1f.1
ionnex netfs: Mounting other filesystems: succeeded
ionnex kernel: Transparent bridge - Intel Corp. 82801BA/CA/DB PCI Bridge
ionnex kernel: PCI: Using IRQ router PIIX [8086/24c0] at 00:1f.0
ionnex kernel: PCI: Found IRQ 10 for device 00:1f.1
ionnex kernel: PCI: Sharing IRQ 10 with 00:1d.2
ionnex apmd[3397]: Version 3.0.2 (APM BIOS 1.2, Linux driver 1.16)
ionnex apmd: apmd startup succeeded
ionnex kernel: PCI: Sharing IRQ 10 with 01:00.1
ionnex kernel: isapnp: Scanning for PnP cards...
ionnex kernel: isapnp: No Plug & Play device found
ionnex kernel: Linux NET4.0 for Linux 2.4
ionnex kernel: Based upon Swansea University Computer Society NET3.039
ionnex autofs: automount startup succeeded
ionnex kernel: Initializing RT netlink socket
ionnex kernel: apm: BIOS version 1.2 Flags 0x07 (Driver version 1.16)
ionnex kernel: Starting kswapd
ionnex kernel: allocated 32 pages and 32 bhs reserved for the highmem bounces
ionnex kernel: VFS: Disk quotas vdquot_6.5.1
ionnex kernel: Detected PS/2 Mouse Port.
ionnex kernel: pty: 2048 Unix98 ptys configured
ionnex kernel: Serial driver version 5.05c (2001-07-08) with MANY_PORTS MULTIPORT SHARE_IRQ SERIAL_PCI ISAPNP en
abled
ionnex kernel: ttyS0 at 0x03f8 (irq = 4) is a 16550A
ionnex kernel: ttyS1 at 0x02f8 (irq = 3) is a 16550A
ionnex kernel: Real Time Clock Driver v1.10e
ionnex kernel: Floppy drive(s): fd0 is 1.44M
ionnex kernel: FDC 0 is a post-1991 82077
ionnex kernel: NET4: Frame Diverter 0.46
ionnex kernel: RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
ionnex kernel: Uniform Multi-Platform E-IDE driver Revision: 7.00beta3-.2.4
ionnex kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
ionnex kernel: ICH4: IDE controller at PCI slot 00:1f.1
ionnex kernel: PCI: Found IRQ 10 for device 00:1f.1
ionnex kernel: PCI: Sharing IRQ 10 with 00:1d.2
ionnex kernel: PCI: Sharing IRQ 10 with 01:00.1
ionnex kernel: ICH4: chipset revision 2
ionnex kernel: ICH4: not 100%% native mode: will probe irqs later
ionnex kernel: ide0: BM-DMA at 0xcc00-0xcc07, BIOS settings: hda:pio, hdb:pio
ionnex kernel: ide1: BM-DMA at 0xcc08-0xcc0f, BIOS settings: hdc:DMA, hdd:pio
ionnex kernel: hdc: CDU5211, ATAPI CD/DVD-ROM drive
ionnex kernel: ide1 at 0x170-0x177,0x376 on irq 15
ionnex kernel: ide-floppy driver 0.99.newide
ionnex kernel: ide-floppy driver 0.99.newide
ionnex kernel: md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
ionnex kernel: md: Autodetecting RAID arrays.
ionnex kernel: md: autorun ...
ionnex kernel: md: ... autorun DONE.
ionnex kernel: NET4: Linux TCP/IP 1.0 for NET4.0
ionnex kernel: IP Protocols: ICMP, UDP, TCP, IGMP
ionnex kernel: IP: routing cache hash table of 16384 buckets, 128Kbytes
ionnex kernel: TCP: Hash tables configured (established 262144 bind 65536)
ionnex kernel: Linux IP multicast router 0.06 plus PIM-SM
ionnex kernel: NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
ionnex kernel: RAMDISK: Compressed image found at block 0
ionnex kernel: Freeing initrd memory: 241k freed
ionnex kernel: VFS: Mounted root (ext2 filesystem).
ionnex kernel: SCSI subsystem driver Revision: 1.00
ionnex sshd: succeeded
ionnex kernel: PCI: Found IRQ 11 for device 01:00.0
ionnex kernel: PCI: Sharing IRQ 11 with 00:1f.3
ionnex kernel: sym53c8xx: at PCI bus 1, device 0, function 0
ionnex kernel: sym53c8xx: setting PCI_COMMAND_PARITY...(fix-up)
ionnex kernel: sym53c8xx: 53c1010-33 detected with Symbios NVRAM
ionnex kernel: PCI: Found IRQ 10 for device 01:00.1
ionnex kernel: PCI: Sharing IRQ 10 with 00:1d.2
ionnex kernel: PCI: Sharing IRQ 10 with 00:1f.1
ionnex kernel: sym53c8xx: at PCI bus 1, device 0, function 1
ionnex kernel: sym53c8xx: setting PCI_COMMAND_PARITY...(fix-up)
ionnex kernel: sym53c8xx: 53c1010-33 detected with Symbios NVRAM
ionnex kernel: sym53c1010-33-0: rev 0x1 on pci bus 1 device 0 function 0 irq 11
ionnex kernel: sym53c1010-33-0: Symbios format NVRAM, ID 7, Fast-80, Parity Checking
ionnex kernel: sym53c1010-33-0: on-chip RAM at 0xec060000
ionnex kernel: sym53c1010-33-0: restart (scsi reset).
ionnex kernel: sym53c1010-33-0: handling phase mismatch from SCRIPTS.
ionnex kernel: sym53c1010-33-0: Downloading SCSI SCRIPTS.
ionnex kernel: sym53c1010-33-1: rev 0x1 on pci bus 1 device 0 function 1 irq 10
ionnex kernel: sym53c1010-33-1: Symbios format NVRAM, ID 7, Fast-80, Parity Checking
ionnex kernel: sym53c1010-33-1: on-chip RAM at 0xec062000
ionnex kernel: sym53c1010-33-1: restart (scsi reset).
ionnex kernel: sym53c1010-33-1: handling phase mismatch from SCRIPTS.
ionnex kernel: sym53c1010-33-1: Downloading SCSI SCRIPTS.
ionnex kernel: scsi0 : sym53c8xx-1.7.3c-20010512
ionnex kernel: scsi1 : sym53c8xx-1.7.3c-20010512
ionnex kernel: blk: queue c3e33e14, I/O limit 1048575Mb (mask 0xffffffffff)
ionnex kernel: Vendor: SEAGATE Model: ST373307LW Rev: 0007
ionnex kernel: Type: Direct-Access ANSI SCSI revision: 03
ionnex kernel: blk: queue c3e33c14, I/O limit 1048575Mb (mask 0xffffffffff)
ionnex kernel: Vendor: SEAGATE Model: ST373307LW Rev: 0007
ionnex kernel: Type: Direct-Access ANSI SCSI revision: 03
ionnex kernel: blk: queue c3fb5014, I/O limit 1048575Mb (mask 0xffffffffff)
ionnex kernel: sym53c1010-33-0-<0,0>: tagged command queue depth set to 8
ionnex kernel: sym53c1010-33-0-<1,0>: tagged command queue depth set to 8
ionnex kernel: Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
ionnex kernel: Attached scsi disk sdb at scsi0, channel 0, id 1, lun 0
ionnex kernel: sym53c1010-33-0-<0,*>: FAST-80 WIDE SCSI 160.0 MB/s (12.5 ns, offset 62)
ionnex kernel: SCSI device sda: 143374744 512-byte hdwr sectors (73408 MB)
ionnex kernel: Partition check:
ionnex kernel: sda: sda1 sda2 sda3 sda4 < sda5 sda6 >
ionnex kernel: sym53c1010-33-0-<1,*>: FAST-80 WIDE SCSI 160.0 MB/s (12.5 ns, offset 62)
ionnex kernel: SCSI device sdb: 143374744 512-byte hdwr sectors (73408 MB)
ionnex kernel: sdb: sdb1 sdb2 sdb3 sdb4 < sdb5 sdb6 >
ionnex kernel: Journalled Block Device driver loaded
ionnex kernel: EXT3-fs: INFO: recovery required on readonly filesystem.
ionnex kernel: EXT3-fs: write access will be enabled during recovery.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3-fs: sd(8,3): orphan cleanup on readonly fs
ionnex kernel: EXT3-fs: sd(8,3): 6 orphan inodes deleted
ionnex kernel: EXT3-fs: recovery complete.
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: Freeing unused kernel memory: 132k freed
ionnex kernel: usb.c: registered new driver usbdevfs
ionnex kernel: usb.c: registered new driver hub
ionnex kernel: PCI: Found IRQ 5 for device 00:1d.7
ionnex kernel: PCI: Sharing IRQ 5 with 01:06.0
ionnex kernel: ehci-hcd 00:1d.7: Intel Corp. 82801DB USB EHCI Controller
ionnex kernel: ehci-hcd 00:1d.7: irq 5, pci mem f8886000
ionnex kernel: usb.c: new USB bus registered, assigned bus number 1
ionnex kernel: ehci-hcd 00:1d.7: enabled 64bit PCI DMA
ionnex kernel: PCI: 00:1d.7 PCI cache line size set incorrectly (0 bytes) by BIOS/FW.
ionnex kernel: PCI: 00:1d.7 PCI cache line size corrected to 128.
ionnex apmd[3397]: Charge: * * * (-1% unknown)
ionnex kernel: ehci-hcd 00:1d.7: USB 2.0 enabled, EHCI 1.00, driver 2003-Jan-22
ionnex kernel: hub.c: USB hub found
ionnex kernel: hub.c: 6 ports detected
ionnex kernel: usb-uhci.c: $Revision: 1.275 $ time 13:50:45 Dec 18 2003
ionnex kernel: usb-uhci.c: High bandwidth mode enabled
ionnex kernel: PCI: Found IRQ 9 for device 00:1d.0
ionnex kernel: PCI: Sharing IRQ 9 with 00:02.0
ionnex kernel: usb-uhci.c: USB UHCI at I/O 0xb800, IRQ 9
ionnex kernel: usb-uhci.c: Detected 2 ports
ionnex kernel: usb.c: new USB bus registered, assigned bus number 2
ionnex kernel: hub.c: USB hub found
ionnex kernel: hub.c: 2 ports detected
ionnex kernel: PCI: Found IRQ 11 for device 00:1d.1
ionnex kernel: usb-uhci.c: USB UHCI at I/O 0xb000, IRQ 11
ionnex kernel: usb-uhci.c: Detected 2 ports
ionnex network: Bringing up interface eth0: succeeded
ionnex kernel: usb.c: new USB bus registered, assigned bus number 3
ionnex kernel: hub.c: USB hub found
ionnex kernel: hub.c: 2 ports detected
ionnex kernel: PCI: Found IRQ 10 for device 00:1d.2
ionnex kernel: PCI: Sharing IRQ 10 with 00:1f.1
ionnex kernel: PCI: Sharing IRQ 10 with 01:00.1
ionnex kernel: usb-uhci.c: USB UHCI at I/O 0xb400, IRQ 10
ionnex kernel: usb-uhci.c: Detected 2 ports
ionnex kernel: usb.c: new USB bus registered, assigned bus number 4
ionnex kernel: hub.c: USB hub found
ionnex kernel: hub.c: 2 ports detected
ionnex kernel: usb-uhci.c: v1.275:USB Universal Host Controller Interface driver
ionnex kernel: usb.c: registered new driver hiddev
ionnex kernel: usb.c: registered new driver hid
ionnex kernel: hid-core.c: v1.8.1 Andreas Gal, Vojtech Pavlik <vojtech@suse.cz>
ionnex kernel: hid-core.c: USB HID support drivers
ionnex kernel: mice: PS/2 mouse device common for all mice
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,3), internal journal
ionnex kernel: Adding Swap: 2048248k swap-space (priority -1)
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex xinetd[3448]: pmap_set failed. service=sgi_fam program=391002 version=2
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,1), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,5), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,2), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,18), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,17), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,21), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,19), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: kjournald starting. Commit interval 5 seconds
ionnex kernel: EXT3 FS 2.4-0.9.19, 19 August 2002 on sd(8,22), internal journal
ionnex kernel: EXT3-fs: mounted filesystem with ordered data mode.
ionnex kernel: parport0: PC-style at 0x378 (0x778) [PCSPP,TRISTATE,EPP]
ionnex kernel: parport0: irq 7 detected
ionnex kernel: ip_tables: (C) 2000-2002 Netfilter core team
ionnex kernel: Intel(R) PRO/1000 Network Driver - version 5.0.43-k1
ionnex kernel: Copyright (c) 1999-2003 Intel Corporation.
ionnex kernel: PCI: Found IRQ 12 for device 01:05.0
ionnex kernel: eth0: Intel(R) PRO/1000 Network Connection
ionnex kernel: PCI: Found IRQ 5 for device 01:06.0
ionnex kernel: PCI: Sharing IRQ 5 with 00:1d.7
ionnex kernel: eth1: Intel(R) PRO/1000 Network Connection
ionnex kernel: ip_tables: (C) 2000-2002 Netfilter core team
ionnex kernel: e1000: eth0 NIC Link is Up 10 Mbps Full Duplex
ionnex kernel: ip_tables: (C) 2000-2002 Netfilter core team
ionnex xinetd[3448]: xinetd Version 2.3.11 started with libwrap loadavg options compiled in.
ionnex xinetd[3448]: Started working: 1 available service
ionnex xinetd: xinetd startup succeeded
ionnex sendmail: sendmail startup succeeded
ionnex sendmail: sm-client startup succeeded
ionnex gpm: gpm startup succeeded
ionnex crond: crond startup succeeded
ionnex xfs: xfs startup succeeded
ionnex xfs: ignoring font path element /usr/X11R6/lib/X11/fonts/misc:unscaled (unreadable)
ionnex xfs: ignoring font path element /usr/X11R6/lib/X11/fonts/75dpi:unscaled (unreadable)
ionnex xfs: ignoring font path element /usr/X11R6/lib/X11/fonts/100dpi:unscaled (unreadable)
ionnex xfs: ignoring font path element /usr/X11R6/lib/X11/fonts/misc (unreadable)
ionnex xfs: ignoring font path element /usr/X11R6/lib/X11/fonts/Type1 (unreadable)
ionnex xfs: ignoring font path element /usr/X11R6/lib/X11/fonts/Speedo (unreadable)
ionnex xfs: ignoring font path element /usr/X11R6/lib/X11/fonts/cyrillic (unreadable)
ionnex anacron: anacron startup succeeded
ionnex atd: atd startup succeeded
ionnex rhnsd[3555]: Red Hat Network Services Daemon starting up.
ionnex rhnsd: rhnsd startup succeeded
ionnex named[3579]: starting BIND 9.2.3 -c /etc/namedb/named.conf -u named
ionnex named[3579]: using 1 CPU
ionnex named[3579]: loading configuration from '/etc/namedb/named.conf'
ionnex named[3579]: no IPv6 interfaces found
ionnex named[3579]: listening on IPv4 interface eth0, 192.168.1.2#53
ionnex named[3579]: command channel listening on 127.0.0.1#953
ionnex named[3579]: zone 0.0.127.IN-ADDR.ARPA/IN: loaded serial 20040210
ionnex named[3579]: zone 0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.IP6.INT/IN: loaded serial
20040210
ionnex named[3579]: running
ionnex kernel: ip_tables: (C) 2000-2002 Netfilter core team
ionnex kernel: ip_conntrack version 2.1 (8192 buckets, 65536 max) - 292 bytes per conntrack
ionnex rc: Starting webmin: succeeded

MYSQL CONFIG

[client]
port = 3306
socket = /tmp/mysql.sock

# The MySQL server
[mysqld]
user=mysql
basedir=/usr/local/mysql
datadir=/var/db/mysql
port = 3306
socket = /tmp/mysql.sock
pid-file =/var/db/mysql/ion.ion.com.pid
skip-locking
key_buffer = 256M
max_allowed_packet = 1M
table_cache = 256
sort_buffer_size = 1M
read_buffer_size = 1M
myisam_sort_buffer_size = 64M
thread_cache = 8
query_cache_size= 16M
# Try number of CPU's*2 for thread_concurrency
thread_concurrency = 2

skip-innodb

[mysqldump]
quick
max_allowed_packet = 16M

[mysql]
no-auto-rehash
#safe-updates

[isamchk]
key_buffer = 128M
sort_buffer_size = 128M
read_buffer = 2M
write_buffer = 2M

[myisamchk]
key_buffer = 128M
sort_buffer_size = 128M
read_buffer = 2M
write_buffer = 2M

[mysqlhotcopy]
interactive-timeout

## Apache2 ##

>httpd.conf

Timeout 300
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 15
<IfModule worker.c>
StartServers 2
MaxClients 150
MinSpareThreads 25
MaxSpareThreads 75
ThreadsPerChild 25
MaxRequestsPerChild 0
</IfModule>

Server version: Apache/2.0.48
Server built: Feb 10 2004 03:48:49
Server's Module Magic Number: 20020903:4
Architecture: 32-bit
Server compiled with....
-D APACHE_MPM_DIR="server/mpm/worker"
-D APR_HAS_SENDFILE
-D APR_HAS_MMAP
-D APR_HAVE_IPV6 (IPv4-mapped addresses enabled)
-D APR_USE_PROC_PTHREAD_SERIALIZE
-D APR_USE_PTHREAD_SERIALIZE
-D SINGLE_LISTEN_UNSERIALIZED_ACCEPT
-D APR_HAS_OTHER_CHILD
-D AP_HAVE_RELIABLE_PIPED_LOGS
-D HTTPD_ROOT="/usr/local/apache2"
-D SUEXEC_BIN="/usr/local/apache2/bin/suexec"
-D DEFAULT_SCOREBOARD="logs/apache_runtime_status"
-D DEFAULT_ERRORLOG="logs/error_log"
-D AP_TYPES_CONFIG_FILE="conf/mime.types"
-D SERVER_CONFIG_FILE="conf/httpd.conf"


## Apache2+Modperl ##

>httpd.conf

Timeout 300
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 15
<IfModule worker.c>
StartServers 2
MaxClients 120
MinSpareThreads 20
MaxSpareThreads 60
ThreadsPerChild 20
MaxRequestsPerChild 10000
</IfModule>

Server version: Apache/2.0.48
Server built: Feb 12 2004 03:10:20
Server's Module Magic Number: 20020903:4
Architecture: 32-bit
Server compiled with....
-D APACHE_MPM_DIR="server/mpm/worker"
-D APR_HAS_SENDFILE
-D APR_HAS_MMAP
-D APR_HAVE_IPV6 (IPv4-mapped addresses enabled)
-D APR_USE_PROC_PTHREAD_SERIALIZE
-D APR_USE_PTHREAD_SERIALIZE
-D SINGLE_LISTEN_UNSERIALIZED_ACCEPT
-D APR_HAS_OTHER_CHILD
-D AP_HAVE_RELIABLE_PIPED_LOGS
-D HTTPD_ROOT="/usr/local/apache2_perl"
-D SUEXEC_BIN="/usr/local/apache2_perl/bin/suexec"
-D DEFAULT_SCOREBOARD="logs/apache_runtime_status"
-D DEFAULT_ERRORLOG="logs/error_log"
-D AP_TYPES_CONFIG_FILE="conf/mime.types"
-D SERVER_CONFIG_FILE="conf/httpd.conf"

Unfortunately, mod_perl and apache2.0 is still not a good combination. I run some servers with apache1.3 exclusively because of that reason.

Originally posted by optimalix
Unfortunately, mod_perl and apache2.0 is still not a good combination. I run some servers with apache1.3 exclusively because of that reason.

Agree 100%. He can try - compile without multithread support, go with mod_perl2, get back to Apache 1.3. In fact this reason (mod_perl) + handling of connectors for Tomcat makes my choice also in favor of Apache 1.3.

Peter.

Hi guys,

Do i need do some tuning on my httpd.conf and my.cnf ?

And besides apache2+modperl, what else can we do to optimize the performance?

Eg: balancing the load, putting dell1 - webserver, dell2 - apache2+modperl, dell3 for mysql ..? :(