Monday, 16 April 2007

Yahoo! Mail AddressGuard and Facebook Fun

So some of my friends are on Facebook and I was feeling a bit left out. But I didn't want to open my inbox to a potential flood of spam which is what apparently happens sometimes with these social networking sites. This prompted to me to look into the Yahoo! mail AddressGuard feature.

The basic idea is that you create a default name (jrmailgate in my case) and then you can easily setup a number of disposible addresses, so jrmailgate-facebook@ etc. If one address starts getting too much spam, or has found itself onto a bulk list, simply trash the address. It seems like a pretty elegant way to protect your private address.

To be honest, I'm not sure I "get" the idea of a social networking site. Seems like a lot of hassle, but it's all "Web 2.0 goodness" so I should probably make an effort...
:-)

Incidently, as a follow-up to my previous post, Yahoo! Mail does not have the same nifty synchronisation capabilities as Gmail, so I'm still looking for a solution to getting a *complete* backup of my Yahoo! mailbox (including sent items) without losing the ability to access it from anywhere. Ah well, at least it's sunny... :-)

Thursday, 5 April 2007

Backing up Gmail

I've started using Gmail more and more, primarily because I like the ability to read and respond to emails while away from my home machine. The only concern is that a copy of my email remains on Google's servers and there is no guarantee that this will be available.

The ideal way to fix this is to use the POP capability to download a copy of the email for archiving purposes. This is in preference to the mail forwarding option as it will also download the sent email items (the forwarding option will not enable archiving of sent mail).

Once this is done, it's a matter of configuring your email client (Thunderbird in my case) to download from Gmail's POP server. One important thing to remember is to "leave a copy of the email on the server". This will allow access to Gmail through a browser as well as in Thunderbird.

For sending, set the Thunderbird account to send via Gmail's SMTP server. This will magically save a copy of the outgoing email in the Gmail sent email list.

Finally, setup a filter rule in Thunderbird to move all email from xxx@gmail.com (where xxx is your own email address) into the Sent folder. This will keep things nice and tidy in your email client.

Not sure if the same can be done with Yahoo! mail. Will investigate...

Friday, 16 March 2007

ZFS and RAID levels

As a builder of Ingres systems, we tend to use RAID 1+0 (aka RAID10) for most of our systems. RAID5 is not recommended for databases due to the read-parity-write that is required for every write operation. Furthermore, the ability of a mirrored system to read two separate data blocks in parallel can give RAID1 a significant read performance advantage.

With the arrival of ZFS in Solaris 10, I've been reading up a little to see if this accepted wisdom changes with RAIDZ. A very interesting article (http://blogs.sun.com/roch/entry/when_to_and_not_
t o) basically explains that the read potential for a RAIDZ filesystem can be significantly less than that of a RAID10. However, for non-data filesystems (such as the checkpoint backup area), RAIDZ can provide some advantages. In these areas, disk I/O is less critical (the Ingres archiver writing data to the journals is performed asynchronously to a users query execution) and sequential write operations are handled better than random I/O, making it ideal for checkpoints and journals, but less good for data areas.

Will now be putting this theory into practice...

Friday, 2 March 2007

Ingres Performance Tuning

I'm on an Ingres Performance Tuning course this week. Just completed day 4 of 5 and it's all been quite tough. There are many ways to tune a DBMS including choosing your table structures (heap, hash, isam and btree), when to use secondary indexes, how to examine the Query Execution Plan (QEP - the way the DBMS chooses how it will execute a query), key design, application design consideration, which have all led us to the locking and logging system today. The subsystems that enforce consistency but allow concurrency. The number of options and strategies are many and an Ingres installation is extremely flexible (it needs to be when systems have hundreds of concurrent users and databases in the hundreds of gigabytes).

The whole thing is really quite amazing and I bang my head against a wall when people automatically assume "open source database" equals only MySQL or Postgres.

I admit it - I like Ingres - I'm a fan. It's a free download and you can get it up and running (on Linux or Windows) very easily. There is a JDBC driver,
.NET connector, PHP, Python and Perl interfaces. Go on, try something new... :-)

Wednesday, 31 January 2007

New monitor (Iiyama 22" widescreen minireview)

After working with an Iiyama 17" CRT for the last few years, I recently decided to upgrade to a TFT; something that uses a bit less power and takes up a lot less space. I've always thought highly of Iiyama as a manufacturer and decided to get their new 22" widesceen LCD.

It's a nice piece of kit.

The screen is big (obviously) and runs at a pleasant 1680x1050, which although not as high as my work laptop (1980x1600), is very comfortable to work with. It's amazing how much more desktop there is to work with...
:-)

There's not a lot else for me to say about it - colours are bright, resolution good, it's got a black casing which fits in nicely with my black Shuttle next to it. All in all, I'm pleased with this purchase and hope it'll last as long as the trusty (and yellowed) CRT.

Friday, 5 January 2007

Getting Virtual

As someone who has spent the last 5+ years specifying and building servers, the shift to a virtual environment has been... interesting. From the moment I first heard about SAN technologies and, later, virtualisation, it's made a lot of sense. However, it's only in the last month or so, when taking a step back and thinking things through has the true potential hit me.

(Apologies to those who are quicker than I am; I'm sure this will cause a few people to go "duh!").

Previously, when asked to specify a server, a number of metrics would be considered: Processor load, memory requirements, disk usage and performance. These would be totted up and a server specified. The result is a server room with a large number of servers, most of them 1U, sitting idle for the majority of the time while waiting to run a single service. Of course, when it came to adding more disk, this was often a problem as the previously specified server is now full (typically a 1U server would be specified as more would be overkill... until the upgrade is required).

The implementation of a SAN therefore made a lot of sense. By consolidating all storage into a central location, disks can be added and distributed as necessary. The first use of the SAN was quite reactive and basically involved assigning a bunch of disk to a box for file serving. The second server to use the SAN basically took so much disk (24x73GB FC disks) that the effect was the same as adding the arrays directly. The next step however will be more intelligent - to effectively utilise a SAN, I think you need to provision the storage up front so that you can proactively assign it when needed.

The same applies I think to virtualisation. I've spent a couple of weeks now working on the detail of a new server room infrastructure that is virtual. Getting my head around the idea that I no longer need to be specifying specific servers, but rather capacity, has taken some mental adjustment. The result is a breakdown of a server into components that units can purchase, starting with a basic server (1xCPU, 512MB RAM, 1x36GB disk) and then scaling up as necessary.

So today, I plan to take a close look at specifying a couple of servers that are meaty enough to host a number of virtual machines. I was originally looking at the Sun Fire X4100 - a 1U server capable of supporting 2 x dual core Opteron processors. Having spent a bit of time getting to grips with the possibility of creating a large number of servers as necessary, the X4600 (capable of 8 x dual core Opterons) seems more appealing, as additional capacity can be bought at a later date.

All exciting stuff...

Wednesday, 3 January 2007

Synchronising - progress so far

Got back to work this morning and setup Yahoo! Intellisync. Now got it working synchronising my Outlook to my Yahoo Calendar, Address Book and Notepad. All good so far.

Tried yesterday to get Evolution working with my Palm. Managed to get a backup of the Palm using pilot-xfer, and also got sync working with Kpilot, but gnome-pilot just doesn't want to work. It just sits there when I try and sync and eventually times out. Will do some more investigation...