12th of March 2003
There has been some good news for the NoChex service, when my NoChex account for the cardman.com site is now Xtra Enabled.
What this term means is that people who own a UK debit card can now pay me on-line without having to open a NoChex account first. As at the NoChex payment gateway screen you can now enter your debit card details directly and pay me within seconds, even if you have never heard of NoChex before.
This is indeed good news, when this will save new customers having to spend a few days opening an account first, but unfortunately there are a few limitations.
The "secret plot" to all this is that it very much appears like that along with the instant on-line payment NoChex will open you a NoChex account anyway. This is hardly a big deal, but it does go to explain what appears to be a second limitation.
That limitation is that you could well only be able to use that feature just once, where after that you will be expected to pay through your NoChex account. This could cause delays with people placing two orders in a short amount of time, and of course those lacking funds in their NoChex account.
The final and most important limitation is that with your one shot at this new feature, then you are only allowed to pay me up to £50 immediately. More than this and I would expect that you would have to wait for your account to open first, where I also believe that this £50 limit is a bit low, when in my view £100 would have been much better.
To sum all this up, then this new feature is all about NoChex attracting new customers. And the thing I am most wondering is if they charge their usual 99p fee during this transaction, where the "profit" idea says that they will and yet "new customers" say that they will not.
Well maybe one of my new customers can confirm this, when since I already own three NoChex accounts, then I do not feel like opening another one. Anyway, this is good news to new customers owning a UK debit card, when now you can pay me on-line straight away.
The final aspect to all this is that they have not Xtra Enabled my NoChex account for the cardman.co.uk site yet, which means that at the moment you will need to visit my cardman.com site to make use of this feature. As my cardman.co.uk NoChex account receives more payments, where it is the cardman.com NoChex account where I withdraw them to my business bank account (and pay the 1%), then you can see why they upgraded the cardman.com account first.
Anyway, since I have just reminded them that these are two accounts for the same service, then I am sure that the cardman.co.uk site will be Xtra Enabled soon enough.
In the end it is pleasing to see that NoChex for once are taking steps to improve their service, when most of last year was spent adding new limitations (like the £90 account balance) and charges (like the 1%). As now they offer seller accounts removing that £90 account limit and this new feature will certainly help out new customers.
All I can wish for now is that they start supporting Credit Cards as well, when by charging like 3 to 4% on such a transaction, then so could they go back to removing this 1% charge on Debit Cards. What I mean is that in 2001 I could withdraw like £2500 a time at the cost of just 99p, but these days doing that same debit would now cost me all of £25.
Moving on to PayHound, then I receive far less payments through this service and with good reason. In fact due to all the problems with this service, then I simply would not recommend using this service unless the customer in question was intent on a fully free transaction.
To highlight what is wrong with PayHound, then lets compare this service to the NoChex one.
1. PayHound do not offer "seller accounts", which means that even my PayHound
account is limited to £90. This creates the additional problem that
transfers to my PayHound account that would exceed my own £90 account
limit would be DENIED (clearly unacceptable).
Naturally, PayHound does have some good features, like with supporting Credit Cards and not charging anything for the transfers. However, it is clear to see due to the much greater number of received payments through the NoChex service that customers prefer paying NoChex's 99p charge to have their order dispatched as fast as is possible instead of waiting around for days in order for PayHound to get around to it.
So in the end NoChex has taken many steps in the past few months to overcome this £90 account limit, but PayHound has taken no steps to overcome this same account limit. I would be in mind to give up supporting PayHound payments, but then not many months ago I thought the same about NoChex. So in the end we can only hope that PayHound wises up like what NoChex already has.
One thing to note is that PayHound is working on adding support for international cards, where like with the old and now bankrupt MinutePay service I doubt that they will be able to do this. After all these banks are very protective about their card payments, where PayHound offering this for free would destroy the whole notion of merchant accounts. Since banks and payment gateway services make vast sums of money from credit card payments, then there is no way that they would give up this profit.
The real aspect here is that PayHound have not got a grip on what it is that people want, which is simply to be able to do easy and fast payments. So before they think about international support, then they need to get their own home service in order, when this £90 account limit is not easy and waiting 10 days is not fast.
Anyway, to sum all this up, then paying me through the NoChex service is both easy and fast (for a small charge), but PayHound's free service is slow and in general a pain in the behind.
One other problem recently is that my cardman.co.uk hosting messed up few a few days last week, where although the site was still working the e-mail system was out of action. Of course the cardman.com site was working fine, but as there has been six known hosting problems since November (mostly denial to upload), then as my hosting is up for renewal at the end of this month, then very soon I will move the cardman.co.uk site to a new hosting service.
As I broke away their control of my domain name a year ago, when I refused to renew it through them and took the matter up with the main domain service (Nominet), then so can I now switch it easily to a new host within a few days. All I have to do now is to find a good host in the UK (with all those extra features), when at the end of this month my hosting through my current provider will be no more.
Also due to my e-mail system on my cardman.co.uk site being down for a few days, unknown to me, then of course all e-mail to this domain were being bounced back. This of course created problems for on-line orders, which is why all customers sending on-line NoChex payments during this time have now been contacted to find out what this payment is for.
Well that is the recent news, but as I just uploaded the previous news as well, then you may care to read the following section about my new RAID system.
8th of March 2003
Well I finally got my four 180 GB hard drive set up in a RAID configuration over a period of one and a half days this week on my also new Highpoint Rocket RAID 404 controller. Test results were as follows, when I recorded these as they happened.
Installation went well, when all new hardware was detected straight away. However, before trying a RAID configuration I upgraded the card to the latest firmware and drivers, where to my surprise I noticed that the latest firmware supported the RAID 5 (Stripping + Parity) option.
The first RAID configuration I tried along with the test results were as follows...
RAID Configuration: 1+0 (Mirroring + Stripping)
As I found out the expected drive speed for my IBM hard drives before I purchased them, then this 43 MB/S average transfer rate (for the whole available size) was not at all good for a stripped RAID array. And so following this test (that I repeated twice) I decided to confirm the expected speed for my hard disks on their own.
However, before I did that I decided to do a simple write test by copying 7.5 GB of data from my existing hard drive to the RAID 1+0 array. This was of course based on the idea that this was possible due to the read speed of my existing hard drive being faster than this write speed and that the system bus could handle this data rate.
Anyway, since this write test was completed in 8 minutes and 28 seconds, then that gave a write speed of 14.78 MB/S. That though is not quite true, when of course due to the mirroring it had to write the data twice, which meant that just one of these two writes would have gone at slightly over double that speed.
Moving on to the speed test of a single drive, then the results were as follows...
RAID Configuration: None
This highlighted that the expected speed of these hard drives were accurate and that my first RAID test did not work out at all well. This made me then think that I should give the RAID 0 configuration a go, when out of all RAID configurations it is RAID 0 (Stripping) that is the fastest.
RAID Configuration: 0 (Stripping)
As you can see the average transfer rate for the fastest RAID 0 configuration was terrible, which left me wondering how on earth to get this RAID system producing a faster rate than with the individual drive scores. Anyway, since my new controller supported RAID 5, then I decided to give that one a go.
And after waiting a couple of hours for it to configure the drives to RAID 5 (damned parity), then the test results were...
RAID Configuration: 5 (Stripping + Parity)
Although I was not expecting great results from RAID 5 (parity makes the write slow), then this was extra disappointing. And to give it a full spin I decided to format this new RAID 5 array, where since this format took simply ages to reach 1%, then so did I go away and have something to eat.
When I got back the format had got up to like 4%, where I then decided to do some on-line research into how to speed up my RAID array.
All of four hours later armed with knowledge on things to test, then I cancelled this format. After all I had things to test and since it was up to 16% when I cancelled, then so was I not happy about waiting a couple of days for it to complete.
Since research had highlighted that people found that the 16k block size provided the best results, then so did I decide to do a simple two drive RAID 0 test using 16k instead of 64k stripped blocks. And since I had two drives spare, then so did I try another option of the 512k block size.
Testing for these two were as follows...
RAID Configuration: 0 (Stripping)
RAID Configuration: 0 (Stripping)
As now highlighted the 16k block size is the way to go, when the read speed for these two drives is about 20 MB/S more than what a single drive can handle. And since I liked this speed so much, then I ran the test (HDTach) twice more with Average Transfer Rate results of 67143 and 65738 KB/S.
My next thought is that if it has a large 20 MB/S increase using two drives in RAID 0, then lets see how much faster all four drives can go. Test results were as follows...
RAID Configuration: 0 (Stripping)
Now that was both disappointing and not at all logical, where as proved using two drives in parallel is faster than a single drive. And so since using four drives should be faster than with using two, then why are we now back to speed equal to a single drive?
This Highpoint controller sounds like the answer to that question, where like with the 16k block size it seems to like to work best one way and not another. Anyway, since I had tried two and four drives in RAID 0, then so did I decide to try three drives. As the start of this test highlighted speed like the four drive results, then so did I cancel this test before it was complete.
It was at that point I was thinking that I may have to put up with using two RAID 0 arrays, with backing up files between the two. As that so far has been the only option that has given good results, but before doing that I decided to try RAID 1+0 again, but this time with the new 16k block size.
As to RAID 5, then the very slow write time had put that idea on-hold for now. The problem with RAID 5 is that it is a great idea, but even the best hardware money can by cannot handle this parity aspect with great speed. That is the reason why I wanted to try a cheap controller card and doing RAID 5 in the software first, when using the main processor (in my case a 1 GHz Athlon) is bound to be more powerful than what comes on a controller card.
Still since I have yet to try RAID 5 with the 16k block size, then I may yet give that a spin.
Anyway, back to the RAID 1+0 test with the 16k block size, where the test results are as follows...
RAID Configuration: 1+0 (Mirroring + Stripping)
Now you can see why this is the option that I am now using, where apart from the automatic mirroring of the data, then I get a bonus of a 6.6 ms average access time.
Sure using HDTach to find out the sustained transfer rate is not all there is to it, but such scores are a good start. And since I purchased fast drives anyway, then the likes of the 8 MB cache on each one of these drives is bound to make other test results just as nice.
Well time to go and do something useful.
3rd of March 2003
For some reason I keep having to mention that I hate it when the 1st of a new month falls on a weekend, when this means that my site is then only updated to this new month on the following Monday.
Looking further into it highlights how often this has happen recently, when for this month the 1st was a Saturday, last month the 1st was a Saturday as well, the first of January was a public holiday and my business was closed, the 1st of December was a Sunday, then going back to September the 1st was a Sunday.
This means that within the past eight months all of five of those months the 1st fell on a non-working day, but we have just gone through an entire run of four 1st of the months on a non-working day.
Looking into future months then only June and November are going to cause further problems, but I had better check into any public holiday that falls on the first.
Moving on; I see that the Euro is being as strange as what the 1st of the month is, when I recall not that long ago I was wondering if this currency rate was going to fall below 1.5 EUR to 1 GBP. That of course it soon did, but the latest currency rate of 1.4497 EUR to 1 GBP means that it has just fallen below 1.45 EUR to 1 GBP as well.
This is of course good news to my many customers currently paying me in the Euro, where now you have to pay me a lot less Euros for the same items.
For example my ELVIS programmer is currently £45 GBP, where only a few months ago this would have been priced at 72.90 EUR, but as you can see for yourself this huge increase in the Euro currency value means that this item now costs 65.24 EUR. So as you can work out for yourself this item now costs you 7.66 EUR less simply due to the change in the currency rate.
When it comes to items like my wafer cards this strong Euro currency rate is good news, when I already know that I have the lowest wafer card prices in the UK, which means that I could now well have the lowest wafer price in most Euro-zone countries as well if not further.
The point of course is that as I can ship my cards to all these countries for just £1 (or more currently 1.45 Euros), then so I can beat local prices without question.
That I guess goes to explain why my wafer cards and other items have been going off to the likes of Ireland, Germany and Spain, when I have been receiving a lot more Euro currency Cheques recently.
It is good to see that the former name problem is now fixed due to the inclusion of the correct payment name of my Euro (and US Dollar) Order Form, but the only problem I have to solve now is to get these customers to sign their cheques. As since these banks are touchy enough over a missing date, then so would they certainly refuse to cash any cheque not signed.
So like one more order received today a non-signed cheque results in an order going nowhere fast.
Speaking of orders fast on the way to nowhere, then the Customer Messages section on my E-Mail page has had a new addition. Maybe I should rename that to the "name and shame" list, but of course that is only a list of contact problems, where apart from the non-signed Euro Cheque, then another customer only sent £41 to cover a £55.22 (including postage and VAT) ELVIS order.
Normally placing an order should be a simple process, which is certainly true now that I have taken steps to solve common problems. So two of these problems occurred through customers not making use of my Order Forms (and paying me the wrong total) and not signing the cheque is just one of those times when the brain switches off.
Well as I cannot do those sections for them, then I have no solution for those not following my wise advice. As sure enough on this site you are bound to find mention of why my Order Forms need to be used (or at least referred to) and why customers should sign their cheques.
Anyway, since this is no real hassle to me, then these customers are only delaying the receipt of their own desired orders.
Since I have been very busy recently, then to start off older news it is the case that some time ago I restocked on the FUN4 cards. These turned up just one day after the last lot expired, which means that I have now got loads of them.
That also means that I have (nearly) emptied my Euro account buying all these wafer cards, where as mentioned before I need your help to refill it. So Euro Cash, Euro Cheques (the signed type preferred!) and Euro bank transfers (SWIFT / CHAPS) are all options in helping me do so. The Euro Cheque collection I currently have is a good start, when I will soon be sending these off, but I will certainly be needing more Euros (hey i'm poor) if I am going to by these FUN6 cards in Euros in a few weeks time.
Some good news is that my long overdue VAT returns are now no longer a problem, where as all of them are now in, then HM C&E now owe me an almost £1700 VAT refund. Those last two VAT returns should have been in at the end of January, but as No3 quit without prior warning on the 14th of January (in breach of a 12-month employment contract), then so doing her work means that these got delayed into February.
And these latest returns (out of six that were overdue) will teach them a lesson, when at no point during these accounts problems have I ever owed HM C&E a debt, when during the past year they have always owed me money.
So in the end they said that I owed them £27,594 with their highly inaccurate estimates, when in the end they owed me exactly £1681.28. This means that their estimates were out by 344%, which I for one happen to find unacceptable.
In their latest and final estimate they said that I owed them another £8125 for this last quarter in a new record estimate, where instead the correct amount was actually that they owed me another £1324.02, which makes their estimate almost £10,000 out in one quarter.
Well anyway as my VAT problems with HM C&E fantasy VAT estimates are now over, then I will be looking forwards to my £1681.28 refund. And should they happen to not pay up, then HM C&E can speak to my own debt management unit.
Now I have just to sort out the Inland Revenue and their Income Tax returns, where I have already seen that quite a lot more data needs to be assembled in order to complete these. The Inland Revenue wants them by tomorrow (yes tomorrow), but that simply is not going to happen.
Well if they keep hasseling me for these tax returns, then I will just quickly fill them in with the data I have and write on them "The data in these forms are completely inaccurate and should be used for reference purposes only". They of course would not like that, when you need to sign these things to say that they are accurate and not inaccurate.
And of course if I did that they would just send them back to me with a note saying to now do them correctly, which is my current plain at this time, where I now need to assemble more data to come up with an accurate total.
These tax people of course lack understanding that as my accounts were taken out of my ownership for 7 months, then so you cannot simply wave a magic wand to make them completely up-to-date again. Due to various problems it has taken another 9 months to complete the entry of now 16 months worth of data and to change my accounts to handle every change in my business during all this time. And as I have seen there are a few small aspects that still need to be completed even now, which is one of the things that I am doing at this time.
So as I say to these people that if they want accurate figures, then so is there a point in the future where this will occur and not before. And so giving me four days to complete tax returns that at minimum will take no less than two weeks of solid work is unhelpful, where in truth this is more likely to take between one and two months.
Well of course they can fine my £60 per day if they need to, but that would just go to put the Maidstone Fraud Squad on to my s**t list. What I mean is that as one section of the state (the Police) is partly responsible for this accounts mess, then if another part of the state (the Inland Revenue) want fine me large fees, then so will the state be paying me back those fees with interest.
Currently it is only FACT I will sure as anything sue over this, but if the state wants to annoy me further over this, then so will the state soon be in court as well. They fine me and so I sue them over fines that they made me have to pay in the first place.
Anyway, before I get annoyed once again over the mess that was created all that long ago on the 6th of November 2001, then to begin with I am off to find a total for all these transaction charges that NoChex have been charging me. My accounts do not yet contain them yet after all, where it was mostly the case that long ago when they were only charging me 99p every two weeks I was not concerned. However, since I must be paying them a few hundred pounds each year now, then this is one expense that needs to be recorded for tax purpose.
And just maybe I can get a VAT refund on those charges as well, when after all this company must be VAT registered.
Well concerning my accounts, then the good news is that this problem is at last coming to an end, even if future problems is not fully clear.
What I mean is that long ago I had No3(a) work for me for just one day before No3(b) became employed, took a lot of time off and then quit. And the news here is that No3(a) will be doing some accounts work for me this coming Saturday.
As it took Mary (No3b) many weeks to start to get a handle on my complex accounts, then one day a week work by No3a (my cousin) is not too promising. Still once the data has been entered, then at least it does not take me long to debug it (as in like one month's worth for each two days of solid work).
Anyway, since No3a can help out with returns as well, then that is most of the weak areas beyond e-mail now solved. E-Mail is still going quite well mind you, where it is just that this is taking over most other things that I have to do. No solution to that problem yet I guess, which means that accounts and taxes work will be on the slow for the known future.
What I am doing tomorrow is to get my server back into full operation, when about a couple of weeks ago one of my two 120GB HDDs in this server failed. As I lost about 80 GB of non-vital media files (thankfully not my Charmed episodes) due to this, then so did I decide that it was time for a RAID system. And for those that do not know RAID is all about using at least two HDDs to backup data and hopefully increase speed at the same time. Reading from two HDDs is after all faster than just reading from one.
So after some research in the matter then using four HDDs in a RAID 5 system seemed the ideal answer, where RAID 5 is about using stripping to speed up drive speed, but then using parity to protect data. In other words if one drive failed, then so can the data be restored from the other drives data and parity. And of course with the hot swap option you can pull out the faulty HDD and put in a replacement while it is still running, where this new drive is restored with the missing data.
The only problem here is that calculating the parity and writing both data + parity to these HDDs is slow, which means that I will have to do some testing between RAID 5 and RAID 0+1 in order to find out what it best. As if RAID 5 is too slow, then so will I go with RAID 0+1, but if the speed is acceptable, then so RAID 5 it is.
Why I do not desire RAID 0+1 to begin with is that this is the stripping + mirror option, which at the end of the day mean that two of the HDDs are simply a mirrored clone of the other two. And as you can figure out for yourself this causes loss of half the available HDD capacity, where in the other hand RAID 5's method of 1 + 1 + 1 = 3 means that only one of the four drives are lost to parity.
Since tomorrow will arrive my four new 180 GB HDD, then the end RAID 5 drive would appear as 540 GB, but under RAID 0+1 this means that the drive will be 360 GB in size. Sure using all 720 GB capacity would be desirable, but then doing so would result in data loss if one fails.
I wrote off the idea of expensive tape backup, when no tape is that size meaning lots of tapes and I could spend this money on live capacity instead.
Why I decided on the drives I got (IBM 180GXP 180Gb 7200rpm ATA100 8mb Cache) before all others was that these drives still come with a three year warranty, where as of late all manufacturers have been reducing their warranties to just one year. These IBM drives are considered fast (faster than my replacement Maxtor), where drives at 7200 RPM and with 8 MB of Cache are ideal.
When it comes to the warranty, then like other manufacturers, even IBM on these same drives yet with 2 MB of Cache, only provide a 12-month warranty. What is more these drives had one of the lowest prices in cost per GB around, which made them the clear choice.
To do RAID I of course needed an extra controller card, where the Highpoint Rocket RAID 404 will be here tomorrow as well. To do RAID 5 I will of course be using Windows 2000 Server, which means that if RAID 5 does work out I may upgrade later to a hardware version if I discover that it is worth the cost.
Also as having five or six HDDs in my server (I have not decided what to do with the 120GB replacement yet) will be creating lots of heat, then I am fortunate that months ago I upgraded the server to use a Lian-Li Midi-Tower case. As my case has space for no less than 7 HDDs (I may be able to squeeze in 10), but of course it comes with four fans blowing air in past the HDDs and out the top and back.
In fact due to this "past the new RAID system" section I decided to replace all my floppy and hard drive cables at the same time, when these flat cables are sure good in blocking air flow. And so tomorrow will arrive the new rounded type, which I am sure will greatly improve air flow.
Also I will have to remember to remove the air input filter, where obtaining maximum air flow means removing the device to keep dust out.
Since my main HDDs in the server runs the O/S, then in order to keep the server ticking on 24 hours a day 365 days a year, then I am thinking of using this replacement 120 GB HDD to put this on either a RAID 1 or RAID 0+1 system. The main problem is that I doubt that I can turn this existing HDD into one drive in a two RAID drive system without having to reformat this drive first.
So in this case I may just leave it until the O/S needs replacing or this drive fails, when I can swap this over to RAID when I need to reinstall the O/S.
Well this is all about keeping my main access point to the Internet working, when my cable modem plugs directly into the server. As if the server fails then loss of Internet access for an Internet based business would be a serious problem. Already I have had one of the two drives in this machine fail, where this time I was fortunate that it was the one that did not contain the important data like the Operating System, the main (hard encrypted) business files and my Charmed Episodes.
I will soon be moving two of those three aspects to the new RAID system, even if technically my business files are already backed up (at least twice). The O/S as mentioned will have to go on and suffer, when I see no easy route of converting this to RAID.
That is the point of why I am undecided as to what to do with this replacement HDD, where waiting for this O/S to fail means that I could use this drive in one of my workstations (here = 1 server, 4 workstations and 2 laptops connected by wireless) instead. As when it does fail one extra day waiting for the (now two) new HDDs to turn up is not that serious, when it would take another day or two to have the Internet access working again and a good week to have things back as they should be.
And of course I do have dail up access to at least report that my server had died if needed.
Well as you can tell I will be busy tomorrow seeing how fast different RAID systems run and making sure that all four of these 180 GB HDDs are what they should be, where speed will be important when I later upgrade to Gigabit Ethernet. After all no single HDD can fill that network bandwidth, but put three HDDs in a RAID system and it then can.
And if I ever get bored with RAID and decide on the backup route instead (beyond already backed up vital data that is), then so does the four channels on my new controller support up to 8 HDDs. I don't need that kind of capacity yet, when even 360 and 540 Gb can be considered an overkill, but maybe one day I will.
Then maybe one day in the far future I will even make a few video clips, like with "how to use your ELVIS programmer".
Anyway, since I have been busy these past couple of weeks, then this long posting today goes and fills things in.
And the two things I forget to say when reading this back is that I just changed the prices of the ATMEGA cards, where the final thing is that the Order Form will no longer automatically open in a new window, when too many people were overlooking the Euro and US Dollar versions if not the entire Order Form Details page.
So now you will have to pick what one of these three Order Forms you need, where this change will more clearly point things out.