Posts

Social Distancing Movie Marathon - Charade

In the UK, we're in the midst of dealing with Covid-19, and although I'm not directly impacted I am spending a lot more time at home what with the cinemas and theatres (my favourite places!) being closed, and my employer has moved to full-time working from home. On one hand this is great! No commuting, no annoying colleagues, no noisy office! Excellent. On the other, the only person I'm likely to see much of for a while is my partner, who is also working from home full-time. We're lucky that we are able to work comfortably in separate rooms, so we only see each other for coffee breaks and at lunchtime. As a consequence, I have found myself watching more movies. And last night I came across this list from the British Film Institute (BFI) - 100 great films available to stream on Netflix and Prime (though they do cheat a little and include BFI Player):  https://www.bfi.org.uk/news-opinion/news-bfi/lists/100-great-films-streaming-netflix-amazon-prime I'm going to

Now certified! DB1-C01 - AWS Certified Database - Specialty

I finally got the email through from the AWS Training group that I had passed the DB-C01 exam and I am now officially Certified! Hurrah! It took 2 months to the day to get the results through. I'm very pleased with the result and am looking forward to getting to put some of this knowledge into practise!

DB1-C01 - AWS Certified Database - Specialty Beta Exam

The beta exam consists of 85 questions and 230 minutes, including 5 minutes pre- and post- for some questions. There's not much information out there at the moment about this exam other than this guide . I completed the exam today, and out of the 230 minutes available I used about 130m (2h 10m) to go through all 85 questions and double-check my answers on questions I flagged. I could have spent longer going over some answers but after 2 hours+ in a room with a noisy air conditioning unit, I was struggling. Sadly, results aren't available immediately for beta exams, with results being made available around 90 days after the end of the beta period (which is today!), so I've no idea how well I did. What I can talk about is what the exam was made up of. It was not as bad as I expected it to be, given that I had all of 3 days to prepare for this (I asked my manager to approve the expense in early December, they only got back to me this week and the beta period ends today!)

Films - a brief review of 2019 and 2020 Goals

In late 2017 thanks to some acquisitions, I landed a new job based in London - still an Oracle DBA but now within a corporate environment. The move to London has meant that I now have access to all the entertainment that London can provide. In 2018 I saw 52 shows (theatre, musicals, etc), and in 2019 I saw 58. The point of this blog post is that also in 2019 I saw over 110 events at the cinema. I phrase that carefully because it wasn’t all films. There were other events too, such as Mark Kermode in 3D, launches, previews, and stuff like that. I’ve decided a platform to write up some reviews of the things I’m seeing would be a good thing, so here we are. I hope to write up some reviews of some of the more memorable films I saw in 2019 and continue writing reviews through stuff I see in 2020: I’d like to keep this going, maybe once a week, but we’ll see. Consistency is not my strongest point.

APEX, SERT, and EPG

APEX, SERT, and EPG Application Express, Security Evaluation and Recommendation Tool, and the Embedded PL/SQL Gateway My organisation has only just started seriously working with APEX and have recently launched our first internal production app (that was a challenge... I was told that this app needed to be in production in 2 days and nobody had bothered to check with me what was needed to make this happen).  I've learned a lot about APEX in the past few weeks and I've got a lot more to go. I recently came across SERT ( https://github.com/OraOpenSource/apex-sert ) and it sounds like a really nifty tool.  Today I found some time to try it out and see if our dev team might find it useful.  I decided to install it in our dev database.  This uses the EPG rather than ORDS. The install didn't go so well.  The first install failed near the end because I think I used a lowercase 'y' for a response somewhere and it got confused, bombing out with an ORA-06502 er

Logon triggers

I've wanted to do this for a long time, but now a migration project to both a new data centre, and a new version of Oracle means I have opportunity to do it and hopefully, do it right. The current database design is not perfect.  There is one application user which is used by apps, users, support and development teams.  At the moment, there is no way to separate traffic, prevent connections from people to the wrong place or from unexpected places, or to do resource limiting.  I can't do any sensible auditing. This changes as of the migration to 12c.  I have defined 7 different services, which will all have different sets of expected users.  After logon triggers will be set up so that all connections to the controlled services have the connection source IP and database username checked.  If either isn't appropriate for the service, the connection attempt is logged and the connection terminated. CREATE OR REPLACE TRIGGER SYSTEM.check_service_appropriate AFTER LOGON ON D

Lies, damned lies, and statistics - a practical demonstration of how table statistics make or break performance

I recently discovered a process that I didn't recognise which had been running for nearly an hour.  With my current employer, the fact that it was a job I didn't recognise wasn't the bit that phased me (I've got used to that kind of thing happening here), but the fact that it had been running for so long was worrying. I checked with the dev team responsible for it, and they told me that it "ran in 15m in development", which is a vastly underpowered VM compared to the physical production server, so the fact that it had been running an hour was cause for concern. I checked the process was actually running and doing work, I checked that it wasn't blocking anything else, and that it looked like it would (probably, eventually) finish at some point. I checked the execution plan (ugh, it was nasty) and compared it to the one on the development server.  The two were completely different.  This prompted me to take a look at the table statistics, and there wer