Somewhere out there!

I just installed Nexenta and I’ve been playing around with zfs. I must say the basic replication system is very cool. Seems to work for zvols as well. Remotely replicated iscsi volumes and nfs data sets, is going to make my backups trivial.

root@hzsilo:/tank/test# ls -l /tank/test2/
total 104
-rw-r--r-- 1 root root 104000 Feb 24 00:54 random
-rw-r--r-- 1 root root 104857600 Feb 24 01:00 random2

root@hzsilo:/tank/test# rm random2
root@hzsilo:/tank/test# zfs snapshot tank/test@10:00pm

root@hzsilo:/tank/test# zfs send -i tank/test@9:59pm tank/test@10:00pm | zfs receive tank/test2@10:00pm
cannot receive: destination has been modified since most recent snapshot

root@hzsilo:/tank/test# zfs rollback -r tank/test2@9:59pm
root@hzsilo:/tank/test# zfs send -i tank/test@9:59pm tank/test@10:00pm | zfs receive tank/test2@10:00pm
root@hzsilo:/tank/test# ls -l /tank/test2/
total 103
-rw-r–r– 1 root root 104000 Feb 24 00:54 random

For those of you who might have missed it, the send receive pipe could have a ‘ssh -C’ part as well. Brilliant. Plus because of the snapshots the deleted data is still there. The whole rollback thing is probably because of atime from the ls.

Some other stats from Wikipedia:

• 2⁴⁸ — Number of snapshots in any file system (2 × 10¹⁴)
• 2⁴⁸ — Number of files in any individual file system (2 × 10¹⁴)
• 16 exabytes (2⁶⁴ byte) — Maximum size of a file system
• 16 exabytes (2⁶⁴ byte) — Maximum size of a single file
• 16 exabytes (2⁶⁴ byte) — Maximum size of any attribute
• 3 × 10²³ petabytes — Maximum size of any zpool
• 2⁵⁶ — Number of attributes of a file (actually constrained to 2⁴⁸ for the number of files in a ZFS file system)
• 2⁵⁶ — Number of files in a directory (actually constrained to 2⁴⁸ for the number of files in a ZFS file system)
• 2⁶⁴ — Number of devices in any zpool
• 2⁶⁴ — Number of zpools in a system
• 2⁶⁴ — Number of file systems in a zpool

With that many snapshots possible you could almost have continue data protection. Even with 100 objects to snapshot at one per mintue; 144,000 snapshots, per day, it will be a long time before any limits are reached.

Technorati Tags: solaris, zfs

This is quite clever. Using flash to build a zero-install voip  phone. No doubt we’ll be seeing phones this on corporate helpdesk pages in due course.

From: jkOnTheRun: Gizmo gets Flashed, makes calls from the browser

Technorati Tags: flash, voip, web

From jkOnTheRun: Polymer Vision to offer rollable eBook reader. One of these will certainly make it easily to give an ebook to my mother.

Technorati Tags: ebooks, hardware

A little rant in following from: Red Hat vs. Microsoft: Who will win

The question I ask is why does Red Hat have to be as large as Microsoft?

Open Source allows everyone to share from each others work. It understands the implicit contract in computing that value creation in information is only occurs when it shared, a version of Metcalfe’s Law. In constrasted Closed Source understands that smaller groups controlling the flow and access to information and innovation allow these groups greater potential wealth.

After all how much would the network cost now, if we were running Microsoft or similar on our routers, servers and phones. Without Linux or BSD the freedom for choice, the ability to choice who innovates and thus allowing the market to decide who is better is removed.

So the innovation that is occurring and been driven in IT at the moment is because of OSS.

A clear example of this is Vmware, Xen and rPath.  Before the advent of Xen a company like rPath would likely have less reason to exist. There would be no software appliances. Vmware would still be printing money for its stock-holders. And for the man in the tench the vision of deploying a enterprise grade mail system like Zimbra in 30 minutes on a rPath platform would be hopeless.

It is not to say that I think huge companies are bad, but that freedom of choice it more vital. I think in the end on the balance of things, it is US all who will win.

Jonathan is a very interesting guy, who seems to have a good vision for Sun. Its good to see an important tech company being lead by a visionary geek business leader, rather than just a business leader. Check out his blog as well.

Setting up and Running Amazon EC2 from Windows. Check it out, very cool.

Especially when you can do this sort of stuff: Mux – Video Transcoding Powered by EC2.

Technorati Tags: ec2, aws

Will have to trying Fake MX records and some of the other tricks on OtherTricks – Spamassassin Wiki at some point.

Technorati Tags: spam, tricks

This has been sitting forgotten on the draft queue for a while. I’m not using this setup at the moment, but the information is still useful.

AoE [1], [2] is a recent protocol developed by Coraid.

ATA over Ethernet is a network protocol registered with the IEEE as Ethernet protocol 0x88a2. AoE is low level, much simpler than TCP/IP or even IP. TCP/IP and IP are necessary for the reliable transmission of data over the Internet, but the computer has to work harder to handle the complexity they introduce.

Users of iSCSI have noticed this issue with TCP/IP. iSCSI is a way to send I/O over TCP/IP, so that inexpensive Ethernet equipment may be used instead of Fibre Channel equipment. Many iSCSI users have started buying TCP offload engines (TOE). These TOE cards are expensive, but they remove the burden of doing TCP/IP from the machines using iSCSI.

…

An alternative to iSCSI, the AoE specification is 8 pages compared with iSCSI’s 257 pages.

The storage hardware sold by Coraid is very cost effective. The basic 1U four SATA disk chasis SR420 is 2000USD, and with the addition of four 750b SATA disks (say about $450) provides 3Tb raw for under 4000 USD in a 1U with RAID 0,1,5,10 or JBOD. This can be combined with their NAS Gateway the CLN20, to provide a reasonable local network storage system.

With Hitachi’s forthcoming 1TB disk, you get 4TB raw. Crazy!

Of course SATA is not the best for database system but if you have a low access archive type system, or a video silo then this is going to work well.

Technorati Tags: aoe, storage, hardware

I’ve been looking at various storage options recently. DRBD is pretty cool:

And, of course, we only will resynchronize those parts of the device that actually have been changed. DRBD has always done intelligent resynchronization when possible. Starting with the DBRD-0.7 series, you can define an “active set” of a certain size. This makes it possible to have a total resync time of 1–3 min, regardless of device size (currently up to 4TB), even after a hard crash of an active node.

DRBD is a network block level intelligent replication protocol. It syncs local disk blocks on two separate nodes and smartly manages dirty blocks while out of communication. Normally this is used for Active/Passive or Active/Active cluster HA. It could also provide a method for off site replication.

Run a storage server with iSCSI or AoE targets on LVM devices over the DRDb block device, with the secondary node on a remote server via a vpn. Specific details about this setup are discussed in Disaster Recovery with “Tele-DRBD” on the Linux HA Wiki. Some details on LVM over DRBD. Also worth checking the NFS on DRBD page.

Technorati Tags: drbd, ha, replication, storage