| Author |
Message |
admin
Admin
|
# Posted: 27 Oct 2007 02:00 - Edited by: admin
Reply
Sementara tgh relaks 2 hari nie...tetiba dapat idea nak try buat noguix cluster edition. Selepas melakukan kajian pasal cluster, requirement, ikut tutorial, macam boleh rasanya dimuatkan dalam linux noguix.
APA ITU CLUSTER?
Nak explain pun aku x pandai. Tapi macam nie la kira nya. Cluster resource contoh web server tengah run di node1, tetiba node1 down, web server tu akan fail over ke node 2 secara automatik. lebih kurang cam tu la.
APA KITA KENE BUAT?
Kita create :
- 2x node (server) + web server cluster resource
- configure node1 + copy configuration ke node2
- start cluster
- test
So, atas dasar cuba-cuba, mungkin noguix cluster edition cuma ada web server sahaja. Kalau ok, baru try masuk mysql & sedmail.
BILA NAK START?
no idea!! x tau lagi..bila aku free la..
ariyossss
|
admin
Admin
|
# Posted: 27 Oct 2007 21:06 - Edited by: admin
Reply
OK... SEBELUM BERMULA.. aku kene buat research dulu bahan2 yg diperlukan :
BAHAN-BAHAN :
* At least two computers. You don't need to have identical hardware in both machines (or amount of memory, kernel, etc.).
* One or more media to send the heartbeat packages (serial, USB or Ethernet cable(s)).
* A Linux OS and the media between the 2 nodes configured and working.
* vmware
Nak mudahkan kerja, aku guna vmware create 2x virtual machine dan running noguix di dalam dua2 node.
PERSEDIAAN :
* The primary node: hostname = node1, ipaddr = 192.168.0.3
* The slave node: hostname = node2, ipaddr = 192.168.0.10
* The IP Apache listens to: 192.168.0.5
VMWARE CONFIGURATION
* create 2x node = node1 & node2
* assign node1 & node2 running from iso
* node1 = use bridged connection (for download purpose from internet if needed)
* node2 = host-only (private network)
NODE1
* change /etc/hosts -> enable ftp download from minux site
* get samba.mo --> install (File sharing bewteen node1, node2 & physical node(winxp)
* start samba client
HEARTBEAT TESTING - STABILITY, RELIABILITY
Selepas diinstall samba pada node1, download file heartbeat.tgz dari winxp. Heartbeat memerlukan "libnet" yg mana sudah terdapat di dalam noguix.
* Download Heartbeat --> convert --> tgz -->convert-->.mo
* Install Hearbeat ke dalam dua-dua nodes.
* configure node1
* change hostname to node1
* masukkan ipaddr node2 ke /etc/hosts
* copy sample basic configuration file dari /usr/doc/heartbeat-2.0.8 (ha_logd.cf, authkeys, ha.cf) to /etc
* edit authkeys,
* edit ha.cf - use_logd yes, bcast eth0, node node1 node2, crm on
* edit ha_logd - logfacility daemon
* copy configuration dari node1 ke node2
* configure node2
* change hostname to node2
* masukkan ipaddr node1 ke /etc/hosts
* NODE1 Start heartbeat --> /etc/rc.d/rc.heartbeat start
Failed! - no ha.cf found in /etc/ha.d/ha.cf
Notes : Much errors bcoz of :
- uid, gid
- logfile, debug
- authkeys file permission must 600
- authkeys use crm method first
- start heartbeat ---> OK! running
* NODE2 Start heartbeat --> /etc/rc.d/rc.heartbeat start
Notes : Much errors bcoz of :
- uid, gid
- logfile, debug
- authkeys file permission must 600
- authkeys use crm method first
- start heartbeat ---> OK! running
Kedua-dua node berfungsi dengan baik secara virtual (vmware)
node1 cuba memperbaharui status node2
node2 cuba memperbaharui status node1 setelah node1 restart
STATUS :
Install HA Cluster = OK!
Running HA = OK!
Create Resource = not yet!
To be continued!
|
admin
Admin
|
# Posted: 31 Oct 2007 06:34 - Edited by: admin
Reply
OK TESTING II
Objektif :
Kita nak pastikan IP address 192.168.0.9 akan sentiasa up dan boleh di ping w/pun primary node ( node1) down.
Scenario
1) Cluster resource ipaddr 192.168.0.9 by default running on node1. Bila shutdown node1, automatik node2 pula akan takeover ipaddr 192.168.0.9.
2) Dan apabila node1 restart dan up semula, node2 akan memulangkan semula ipaddr 192.168.0.9
 
|
admin
Admin
|
# Posted: 1 Nov 2007 20:41 - Edited by: admin
Reply
How to check cluster running or not?
1- /var/log/messages
2- /var/run/heartbeat/rsctmp/IPaddr
3- /etc/ha.d/heartbeat/resource.d/IPaddr <ip-address> status
|
admin
Admin
|
# Posted: 1 Nov 2007 21:52 - Edited by: admin
Reply
TESTING CLUSTER FOR SSH
Steps :
- create /etc/haresources
- create ssh resource.d script
Failed!
notes : comment /etc/rc.d/rc.local hostname noguix
|
admin
Admin
|
# Posted: 29 Nov 2007 11:51
Reply
ok, update noguix cluster edition :
- replace postfix with sendmail
- replace monkey with apache
- add mysql
- DNS & DHCP server include
- ssh & vsftp server include
- Cluster include
- drbd include
- Webmin 1.35 include cluster monitor
Cluster Testing :
- Apache - Active, passive, Fail over, get resource - OK!
- Mysql - Active, active, fail over, get resource - OK!
- Sendmail MTA - active, passive, Fail over, get resource - OK!
|
admin
Admin
|
# Posted: 2 Dec 2007 19:24
Reply
Recent Cluster Update :
- Replaced Webmin-1.35 with Webmin-1.38
- Removed drbd not supported by kernel
- Insert ATA Over Ethernet Module
- Insert aoe tools
To do :
- Extract webmin & fix server path (apache, mysql, sendmail, sshd & etc)
|
admin
Admin
|
# Posted: 27 Dec 2007 10:27
Reply
noguix-cluster-1.0.5RC1 ready for download....
go to www.no-x.org
|
admin
Admin
|
# Posted: 13 Jan 2009 01:25 - Edited by: admin
Reply
Hmm lama x update. Since aku ada target baru untuk year 2009 and LInux cluster one of aku punya target. So bermula hari nie, semua project cluster atas linux noguix akan aku record kan kat sini.
BTW, as per last version noguix cluster edition v1.0.5, satu bende yg berharga x dapat aku masukkan iaitu DRBD. Block device yang digunakan untuk replicate data from cluster node to another.
Few known working command & can be run in NOGUIX + Heartbeat-2.0.8
/usr/lib/heartbeat/hb_standby --help (giving up resource)
/usr/lib/heartbeat/hb_takeover --help (takeover resource)
/usr/lib/heartbeat/ResourceManager listkeys node2 (check resource ip currently running on node2)
/usr/lib/heartbeat/req_resource rc.httpd (require http resource)
|
admin
Admin
|
# Posted: 13 Jan 2009 01:33
Reply
Aku bahagikan Linux-HA nie ke beberapa module
Module 1 - hearbeat installation & basic configuration
Module 2 - Simple IP Clustering
Module 3 - drbd as kernel module
Module 4 - drbd configuration
Module 5 - Simple Apache Clustering
Module 6 - Simple Apache Clustering with drbd
To be reserved!
|
admin
Admin
|
# Posted: 13 Jan 2009 12:17 - Edited by: admin
Reply
REQUIREMENT Heartbeat-STABLE-2.1.4.tar.bz2
- libnet
Build package script need some modification such as "PRGRM"=Heartbeat-STABLE"
Build package script need some modification such as "configure"=ConfigureMe"
Comment line 210 in package script (ln -s ../lddirector)
Point to the right document folder (/usr/doc/heartbeat-2.1.4) instead /usr/doc/Heartbeat-STABLE-2.1.4
|
