I am a fan of the previous incarnation of the PlugComputer so I was excited to see that Marvell have unveiled a new PlugComputer dubbed imaginatively "PlugComputer 3.0."
PlugComputer 3.0 Features:
Smaller sleeker design,
More powerful CPU - 2gz Armanda 300 CPU,
120GB 1.8-inch SATA hard drive,
Wifi,
Bluetooth,
10/100/1000 wired Ethernet,
USB 2.0.
512MB of RAM
512MB of Flash memory
I for one would like to see an additional Ethernet port added to increase application flexibility, for some applications where you are using clustered plugs or even for routing, having multiple Ethernet ports is a must.
Even without multiple ethernet ports, these low power consumption devices really could have a place in SME environments, replacing large cumbersome legacy hardware with compact Linux plug servers.
More information about the PlugComputer can be found here
Recital is a rich and versatile product with many ways to do the same thing. Developers usually write code in the way that they are accustomed to without paying much attention to how this will perform in a multi-user environment with large amounts of users and transactions. The best way to optimize Recital applications is to use the built-in tuning capabilities introduced in Recital 10.
Key features of the Recital scripting language include:
What are the key feature of the Recital database?
- High performance database application scripting language
- Modern object-oriented language features
- Easy to learn, easy to use
- Fast, just-in-time compiled
- Develop desktop or web applications
- Cross-platform support
- Extensive built-in functions
- Superb built-in SQL command integration
- Navigational data access for the most demanding applications
DRBD:
DRBD (Distributed Replicated Block Device) forms the storage redundancy portition of a HA cluster setup. Explained in basic terms DRBD provides a means of achieving RAID 1 behavoir over a network, where whole block devices are mirrored accross the network.
To start off you will need 2 indentically sized raw drives or partitions. Many how-to's on the internet assume the use of whole drives, of course this will be better performance, but if you are simply getting familar with the technology you can repartition existing drives to allow for two eqaully sized raw partitions, one on each of the systems you will be using.
There are 3 DRBD replication modes:
• Protocol A: Write I/O is reported as completed as soon as it reached local disk and local TCP send buffer
• Protocol B: Write I/O is reported as completed as soon as it reached local disk and remote TCP buffer cache
• Protocol C: Write I/O is reported as completed as soon as it reached both local and remote disks.
If we were installing the HA cluster on a slow LAN or if the geogrphical seperation of the systems involved was great, then I recommend you opt for asyncronous mirroring (Protocol A) where the notifcation of a completed write operation occurs as soon as the local disk write is performed. This will greatly improve performance.
As we are setting up our HA cluster connected via a fast LAN, we will be using DRBD in fully syncronous mode, protocol C.
Protocol C involves the file system on the active node only being notified that the write operation was finished when the block is written to both disks of the cluster. Protocol C is the most commonly used mode of DRBD.
/etc/drbd.conf
global { usage-count yes; }
common { syncer { rate 10M; } }
resource r0 {
protocol C;
net {
max-buffers 2048;
ko-count 4;
}
on bailey {
device /dev/drbd0;
disk /dev/sda4;
address 192.168.1.125:7789;
meta-disk internal;
}
on giskard {
device /dev/drbd0;
disk /dev/sda3;
address 192.168.1.127:7789;
meta-disk internal;
}
}
drbd.conf explained:
Global section, usage-count. The DRBD project keeps statistics about the usage of DRBD versions. They do this by contacting a HTTP server each time a new DRBD version is installed on a system. This can be disabled by setting usage-count no;.
The common seciton contains configurations inhereted by all resources defined.
Setting the syncronisation rate, this is accoimplished by going to the syncer section and then assigning a value to the rate setting. The syncronisation rate refers to rate in which the data is being mirrored in the background. The best setting for the syncronsation rate is related to the speed of the network with which the DRBD systems are communicating on. 100Mbps ethernet supports around 12MBps, Giggabit ethernet somewhere around 125MBps.
in the configuration above, we have a resource defined as r0, the nodes are configured in the "on" host subsections.
"Device" configures the path of the logical block device that will be created by DRBD
"Disk" configures the block device that will be used to store the data.
"Address" configures the IP address and port number of the host that will hold this DRBD device.
"Meta-disk" configures the location where the metadata about the DRBD device will be stored.
You can set this to internal and DRBD will use the physical block device to store the information, by recording the metadata within the last sections of the disk.
Once you have created your configuration file, you must conduct the following steps on both the nodes.
Create device metadata.
$ drbdadm create-md r0
v08 Magic number not found
Writing meta data...
initialising activity log
NOT initialized bitmap
New drbd meta data block sucessfully created.
success
Attach the backing device.
$ drbdadm attach r0
Set the syncronisation parameters.
$ drbdadm syncer r0
Connect it to the peer.
$ drbdadm connect r0
Run the service.
$ service drbd start
Heartbeat:
Heartbeat provides the IP redundancy and the service HA functionailty.
On the failure of the primary node the VIP is assigned to the secondary node and the services configured to be HA are started on the secondary node.
Heartbeat configuration:
/etc/ha/ha.conf
## /etc/ha.d/ha.cf on node1
## This configuration is to be the same on both machines
## This example is made for version 2, comment out crm if using version 1
// replace the node variables with the names of your nodes.
crm no
keepalive 1
deadtime 5
warntime 3
initdead 20
bcast eth0
auto_failback yes
node bailey
node giskard
/etc/ha.d/authkeys
// The configuration below set authentication off, and encryption off for the authentication of nodes and their packets.
//Note make sure the authkeys file has the correct permisisions chmod 600
## /etc/ha.d/authkeys
auth 1
1 crc
/etc/ha.d/haresources
//192.168.1.40 is the VIP (Virtual IP) assigned to the cluster.
//the "smb" in the configuration line represents the service we wish to make HA
// /devdrbd0 represents the resource name you configured in the drbd.conf
## /etc/ha.d/haresources
## This configuration is to be the same on both nodes
bailey 192.168.1.40 drbddisk Filesystem::/dev/drbd0::/drbdData::ext3 smb
It would appear that gigabit LAN is not! In fact it often runs at the same speed as 100Mbps LAN. Let's look at why exactly.
After configuring your network you can use the ifconfig command to see what speeds the LAN is connected. Even though 1000Mbps is reported by the card, the reality is that the overall throughtput may well be ~100Mpbs. You can try copying a large file using scp to demonstrate this.
As it turns out, in order to use a gigabit LAN you need to use CAT6 cables. CAT5 and CAT5E are not good enough. End result, the ethernet cards throttle back the speed to reduce dropped packets and errors.
You can find a good article here titled Squeeze Your Gigabit NIC for Top Performance. After tuning up the TCP parameters i found that it made no dfifference. The principal reasons behind low gigabit ethernet performance can be summed up as follows.
- Need to use CAT6 cables
- Slow Disk speed
- Limitations of the PCI bus which the gigabit ethernet cards use
You can get an idea about the disk speed using the hdparm command:
Display the disk partitions and choose the main linux partition which has the / filesystem.
# fdisk -l
Then get disk cache and disk read statistics:
# hdparm -Tt /dev/sda0
On my desktop system the sata disk perfomance is a limiting factor. These were the results:
/dev/sda1:
Timing cached reads: 9984 MB in 2.00 seconds = 4996.41 MB/sec
Timing buffered disk reads: 84 MB in 3.13 seconds = 58.49 MB/sec
Well, that equates to a raw disk read speed of 58.49 * 8 = 467Mbps which is half the speed of a gigabit LAN.
So.. NAS storage with lots of memory looks to be the way to go... If you use the right cables!
The best way to learn Recital is to build some applications. The developers of Recital have written a book "Recital Essentials" which you can read here.
// determine how many Recital users are on the system
nusers = pipetostr("ps -ef | grep db.exe | wc -l")
// the click event handler
private function onclick_sourcetree(e:Event):void {
yourTree.editable = false;
}
// the doubleclick event handler
private function ondoubleclick_sourcetree(e:Event):void {
yourTree.editable = true;
yourTree.editedItemPosition = {columnIndex:0, rowIndex:sourceTree.selectedIndex};
}