Category Archives: Computers

Computers

Snowball effect

We recently migrated from an ADSL line to a 10mb BT leased line. We kept the ADSL as a backup. However, I forgot to change the default gateway of our internal DNS servers to the new line (it was on my ‘do it later, after the dust settles’ lists). This wouldn’t normally cause a problem, although it will slow down people’s internet experience a little as they are getting DNS resolutions from an ADSL line rather than a leased line.

However, when the ADSL line went down today, no-one could use the internet. I couldn’t understand what the problem was, as I knew the internet was working (I could ping an IP address, I just couldn’t resolve a domain name).

The first thing I did was check the DNS servers, and this is when I remembered that they were pointing to the ADSL line rather than the leased line. So I changed their default gateways (and rebooted them for good measure). I was convinced this would do the trick, but it didn’t.

At this stage, the pressure starts to mount. A company without internet is a very unproductive company. And I know very little about DNS. I tested the firewalls (both for the ADSL line and the leased line) and they were able to resolve domain names. The ADSL line actually came back online, but this still didn’t fix the problem.

I was out of ideas, so I phoned a friend. Straight away he asked if port 53 was open on the firewall. This is used for DNS. Of course! I quickly saw that it was open on the ADSL line, but not on the leased line. I opened it up, and bingo, the internet returned, and my users were happy again.

The ADSL line actually went down because an air-con froze and flooded the server room floor, and water must have got into the BT socket. I couldn’t make the connection between an air-con failing and a port being open on the firewall. Would I ever have? It’s times like this when I appreciate the benefits of having someone to call on, who can maybe look at things from a more logical perspective. Because this problem started with a wet floor, I wasn’t able to think logically enough about the real cause of the outage.

Office365 woes – upgrading plan from P1 to E3

I needed to setup new e-mail accounts and a  public domain for a remote company within our group. It is only a small business with a handful of employees, so rather than creating VPNs to head office and integrating them into our Exchange infrastructure, I thought it would be a good idea (and fun) to use Office365 instead.

I set-up a trial account for Microsoft’s Small Business plan (plan P1), and attached it to the company’s new external domain by modifying nameservers to point to Microsoft (this means Microsoft handle DNS and ensures that it is all correct).

Everything was working fine. I then decided that actually the E3 plan is more appropriate. I like the fact that it includes Office subscriptions and also telephone support. I find it strange that they call it their “Midsize and Enterprise” plan, since even small businesses like telephone support.

I’ll simply change my trial P1 account to a paid E3 account, I thought. Simple. Except it isn’t. You can’t upgrade from a P1 account to an E3. Even if you were paying for a P1 account and wanted to upgrade. There is no upgrade path. You have to cancel your existing plan, and take out a new plan. And Microsoft warn you must backup up your data and import it into your new account, or face data loss. Also, you can’t simply transfer your domain from one account to another – it takes time, during which you must make alternative arrangements for e-mail etc.

I find it bizarre that I want to take out a more expensive subscription, but Microsoft won’t let me. So instead of going through the hassle of cancelling and starting over, I will simply stick with the inferior plan and enjoy the cost saving.

Being engaged at work

Rick writes a good blog about being engaged at work, after a recent survey found 58% of people claimed that they weren’t engaged. Am I engaged at work? I think so, mostly. But I don’t think a lot of my colleagues are.
Rick writes:

Managers, especially senior ones, with a high degree of control and often a stake in company stock, express surprise that so many people are disengaged from work. What they should really be asking themselves is why people should be engaged at all. …If managers started with the expectation that people would be disengaged unless they did something to engage them, it might make them better managers.

I’ve found this attitude exists in a lot of companies. Senior managers are exasperated with the lack of engagement by their staff, and their staff are exasperated by the lack of encouragement by their managers. Both parties settle into this long term attitude of blaming the other for their predicament, and neither are willing (or can be bothered) to doing something to change things.

There are things that management can do. At a previous company I worked for, the Directors would assemble all staff  in the canteen and give Powerpoint presentations explaining the company’s financial figures for the quarter (“this is how much we sold”, “this was our borrowing” etc etc). This was often bad news, but I believe that rather than depressing staff, they felt included (and hence engaged) in the company, and wanted to work harder to improve the performance in the next quarter. And it also helped when there was good news. Often workers won’t even be aware when a company wins a big contract – they’re more likely to read it in the newspaper than hear from their employer.

I once delegated a big project to one of my staff. At one point the Finance Director, who was my boss, asked me if he thought the project was being handled well. I said that I didn’t think it was, particularly. “Well, you should step in and change things” I was told. But I didn’t want to. Firstly, I may have been wrong in my opinion. I’m not arrogant enough to assume I always know best. And secondly, I didn’t think the project was being handled badly. It was satisfactory. How else would my staff learn to improve if they weren’t allowed to make decisions (and mistakes) of their own? If I’d been asked for advice, I would have given it. But if my staff were keen and confident to take responsibility, it would have been a mistake to clip their wings.  And whilst the end result might have been an inferior project, the next project they handled would be even better, and the one after that better still. I believe giving people responsibility, independence and autonomy is the best way to foster engagement.

Turning off the Blackberry

There’s a fairly dull article in The Guardian (here) about the perils of checking your phone for e-mails, which strikes a chord with me. A few weeks ago I turned off mobile data on my Blackberry, as I’d started to turn into something I hate. I was checking my e-mail in bed as soon as I woke up, and checking it last thing at night before I turn out the lights. I now figure if something terrible has happened, I’ll find out soon enough when I arrive at work. Rather than using my phone, I check my e-mail on my laptop once or twice on a weekend.

I’m also trying to check e-mail less whilst I’m at work. When I used to just be a programmer, before the internet and before e-mail, I would spend most of my working days doing nothing but programming. I now look back on those years as a golden age for my creativity. Sure, I sometimes got distracted and messed about (but always by having a laugh with colleagues, not reading Facebook), but I would have periods “in the zone”, where I could program for several hours without doing anything else.

I’m trying to get back some of my creativity at work, by ditching the multi-tasking, turning off my phone, and keeping Outlook closed. I’m hoping that not only will I get more done, but what I do will be more creative and will add more value to the company. And if I can get more done, I can create a more stable and reliable IT environment for the company. And if the environment is stable and reliable, I have less need to check my e-mail to see if anything bad has happened.

Work shouldn’t be about working hard, it should be about being creative and being focussed. They’re my two key goals at the moment.

As an added benefit, with mobile data turned off, the battery on my Blackberry seems to last forever.

Pixar and photocopiers

There’s a whole world of difference between a relatively young, trendy Californian company like Pixar or Apple and an old, traditional, English manufacturing company like the one I work for. But we can still learn a lot from American hippies and strive to be creative.

Steve Jobs, on discussing his design of the new Pixar headquarters, once said “There’s a temptation in our networked age to think that ideas can be developed by e-mail. That’s crazy. Creativity comes from spontaneous meetings, from random discussions, you run into someone, you ask what they’re doing, you say “wow!” and soon you’re cooking up all sorts of ideas. So we had the Pixar building designed to promote encounters and unplanned collaborations. If a building doesn’t encourage that you’ll lose a lot of innovation and the magic that’s sparked by serendipity. So we designed the building to make people get out of their offices and mingle in the central atrium with people they might not otherwise see.

How do we do that at our work? We don’t have millions of dollars to spend on a shiny new building, that’s for sure. Hopefully there are little things we can do. We have two photocopiers in our main office, one downstairs and one upstairs. We’ve decided to remove the downstairs photocopier. This means that everyone downstairs now has to walk up the stairs and use the photocopier up there. Of course, they all bitch and whine about the extra effort involved in climbing some stairs, but maybe they’ll start talking to the “upstairs people” a bit more? Maybe some good will come from that? It’s a very small step, I’ll admit.

The other thing I don’t like about our office is that most meetings take place in the Conference Room. This is a horrible, stuffy room with a massive table and tired pictures of our products . It’s so formal. I’d love to create a second meeting room and fill it with comfortable sofas, whiteboards and inspirational pictures. Try and encourage a relaxed atmosphere that promotes new ideas and creativity.

 

 

Dealing with Microsoft support

I’ve written about my favourable experiences with Microsoft InTune here. The product is great. Unfortunately, I’ve had more than a few problems with the simple task of buying the damn thing. It started off with an error on my part, as I bought a fresh subscription rather than converting my trial subscription to a paid-for subscription.

To resolve this, I tried to raise a support ticket from the link on the Admin page (http://go.microsoft.com/fwlink/?LinkID=186758) but I received the error “Firefox has detected that the server is redirecting the request for this address in a way that will never complete.”

I logged a post on the Technet Forum, and received this reply from Microsoft:

Best place to open a support request is with MOCP: https://mocp.microsoftonline.com/site/Support.aspx

Instead of doing this, I actually phoned support, and got some really great help from the Customer Service team. They extended my now-expired trial for another month to give me time to convert the trial subscription to a paid-for subscription.

However, this didn’t work. When I tried to buy the trial, I received the error message:

The offer you requested is not available. This offer may have expired or the account you signed-in with may have already used this offer.

I went onto MOCP, and re-opened the support ticket and waited a few days for a reply. I heard nothing, when I logged on to MOCP to see what was happening I saw that the ticket had been marked as ‘Solved’.  I re-opened the support ticket for a second time and after another day or two I eventually got a reply from the helpful chap who I’d previously spoke to on the phone, who wrote: “It seems like your service request was routed through our technical support team and hence the automated email and delay in arriving at my team.

So I was back on track. Only he couldn’t figure out what was causing the error. After a day, he came back with “I have looked into this issue further and it seems I may need to raise a hosting bug to get the issue resolved. Alternatively, you could set up a new trial and purchase Windows Intune as a new subscription to resolve the issue at; http://www.microsoft.com/en-gb/default.aspx.

I selected his first option. This has caused a new ticket to be raised. I received the following from them: “This is an email to confirm that the new SR for this issue is 1162209999, carrying on from SR 110913-009999. I will now close the latter SR and continue to communicate with you in this service request.

So I don’t know if I’m getting closer. I wrote a while back about the benefits of hosting your e-mail with Microsoft through Office365 (read it here). Now, given that all I’ve been trying to do is give Microsoft my credit card details in order to convert a ten user trial account to a full subscription of InTune, would I really want to trust my mission critical e-mail solution to Microsoft? Would I be confident that if something went wrong it would be resolved quickly, simply and painlessly? Would you?

 

The Cloud

Justin Paul makes another excellent blog post about the cloud, and specifically IaaS, which fits in exactly with my plans for the next few years. I’m currently in the process of refreshing our servers and have budgeted around £40k and a lot of pain to do it, including installing a SAN for the first time. But I explained to the Board my confidence that this would be the last server room infrastructure project that I ever carry out (and I’ve carried out quite a few over the last 12 years as an IT manager). The new equipment will be capitalised over four years, and at the end of those four years, IaaS WILL be the replacement.

The main thing holding things back in the UK is British Telecom and the (slow) speed that they are rolling out superfast fibre optic broadband. Pretty much every small business in the UK still relies on old copper lines for its broadband, giving slow access and unreliability with no service level agreements. IaaS fails pretty quickly when a BT engineer accidently cuts through your broadband connection and then tells you it will be three of four days before he can be bothered to fix it.

Talking to IT resellers about the threat to their business models from the cloud, they’re pretty relaxed that it will be years and years before the UK’s creaking infrastructure is replaced by a network suitable for IaaS. I don’t agree. Things always move quicker that IT people predict. I’m giving it three or four years and then I’m signing up.

Talking about Proliants and CPUs and redundant fans and SANs and LFF versus SFF disks already seems out-of-date. It already feels like life has moved on from this. I can already see the future, but am still forced to proceed with the past. However, running vShpere on all this new hardware still feels like the first stage of our move to the cloud. By creating a private cloud now and running it for three or four years, we’ll be in an excellent position to move some or all of this infrastructure into the cloud. It will just become a case of waiting patiently for BT to do their bit.

 

 

 

 

SAN versus Local Storage

Justin Paul, more than anyone, got me turned on to HP’s P2000 SAN. I was ready to sign the cheque, but now I’m having doubts, after the similarly bearded Scott Alan Miller has been extolling the virtues of local storage. Instead of using a SAN, I just bung loads of disks in my Proliants and run VMs from there. Resellers originally sold the P2000 to me on the grounds that my Proliants will one day fail and my SAN will mitigate this by allowing me to use VMware’s High Availability and vMotion. I nodded my head. It’s a No-Brainer they said. But as Scott points out, why do I anticipate my Proliant will fail but assume my SAN won’t – they’re essentially the same bit of kit and are single points of failure. Suddenly, the benefits of a SAN seem much smaller.

I think what the resellers should have said to me in those early meetings is “Your Proliants probably won’t fail. If you can’t live with that small probability then you need some kind of VSA which will cost you. A P2000 won’t help you here.” Assuming we can live with the probability, I’m not sure the extra benefits of a single SAN outweigh the extra cost. I’m confused.

Font stacks

Font stacks are prioritized lists of fonts, defined in the CSS font-family attribute, that the browser will cycle through until it finds a font that is installed on the user’s system. On the website I designed for my company, I used Verdana, Arial, Helvetica, sans-serif. I have no idea why I preferred Verdana.

These are font stacks on some of my most visited websites:
Dropbox: “lucida grande”, “Segoe UI”, arial, verdana, “lucida sans unicode”, tahoma, sans-serif
Microsoft: “Segoe UI, Tahoma, Arial, Verdana, sans-serif”
Google: arial,sans-serif
BBC: verdana,helvetica,arial,sans-serif
Yahoo: arial, helvetica, clean, sans-serif
Facebook: “lucida grande”,tahoma,verdana,arial,sans-serif
Apple: “Lucida Grande”, “Lucida Sans Unicode”, Helvetica, Arial, Verdana, sans-serif
Oracle: arial, helvetica, sans-serif
LinkedIn: Arial, Helvetica, “Nimbus Sans L”, sans-serif
Symantec: arial, helvetica, sans-serif
VMWare: Arial, Helvetica, sans-serif
Flickr: Arial,Helvetica,sans-serif
Rightmove: Arial, Geneva, sans-serif
Amazon: Verdana, Arial, “Helvetica Neue”, Helvetica, sans-serif

Lucida grande is an Apple font, so Facebook and Dropbox are obviously Mac people. Microsoft are keen on their own Segoe UI, but the rest of the world seems to still prefer Arial. I probably went with Verdana because the BBC use it, but I think I’m going to change to Arial. On our Intranet, I might use Segoe UI, although I can’t decide if I prefer it to Arial or not.

Its easy for someone like me to get very anal about fonts.

Offsite backups

Nightly tape backups now fill our LTO-3 tapes, so we urgently need a new backup solution. Disk to disk backups (to something like an on-premise NAS device stored in another area of the building) are great, but we still need to have a weekly backup stored offsite. To date, I haven’t been able to beat tape as a backup medium in terms of reliability, cost, performance and ease of use.

But I’m really reluctant to spend in excess of £2k on a new LTO-5 tape drive. Its serious money for old-fashioned technology.

One option is to do a weekly backup to an internal 1TB SATA drive, and then simply take that drive off-site. I worry about reliability (dropping a drive wouldn’t be too clever). A 1TB Western Digital Caviar Blue internal hard drive costs £27, whereas an HP LTO-4 tape costs £23. In a year, we use 12 monthly tapes, and four weekly tapes – so 16 tapes in all. So using disks would cost £432 versus £368 for tapes. Although I’m not sure 1TB disks will be big enough, and using 2TB disks will add to the cost considerably (at least in the short term until 2TB disks come down in price).

With tapes, we’ll be using Veeam to backup to disk and then Backup Exec to backup to tape (as Veeam doesn’t do tape backups). Going with disk to disk to disk eliminates the need to use Backup Exec, which is a bonus, who wants to use two different pieces of backup software?

I think disks are the answer. I wonder how many other companies do this. On-line backups to the cloud aren’t the answer (just yet). I’ve just run a speed test on our internet connection and got 0.32Mbps upload speed. This won’t change until British Telecom upgrade our exchange. And even when they do, on-line backups will be significantly more expensive.