Cube cart upgrade from v4 to v5

[familychoice]

A client is interested in upgrading his cubecart powered site to v5. Thought I'd better check the forums first and it seems there are a lot of issues with lost data. One customer listed hundreds of database issues after upgrading, and the official staff response was hardly reassuring:

LOL most of the DB inconsistencies really make no difference at all. Others might though.

Just love the use of a LOL when dealing with customer concerns.
http://forums.cubecart.com/topic/43781-database-after-upgrade/

On a more positive note I've set up a site using the free Open Cart and it's been a breath of fresh air.

Anyone had any experiences of upgrading to Cubecart v5?

[sarahA]

Although I don't use CC so am just going by what I've read on that post, to be honest the CC guy is right, and the helpful member (that the OP doesn't seem to appreciate by responding 'Like I say, what do I know', find it a bit rude to be honest).

Most of those inconsistencies are non issues. The auto increments are of no real concern providing the column max value can cope, which it should be able to go to millions fine. The column orders are of no issue at all providing the mysql code has been written properly (and I would imagine, well hope, it has!), the primary key isn't a major problem really, again providing the code is written properly, you don't have to have a primary key, it's just usually set to enforce rules for protection. And the full text setting is again, just an order thing, but has no bearing.

The only bits I'd be concerned with are new columns missing. Additional columns/tables will always be there. It doesn't matter if you have a column called 'widgets' in a specific table, if it's not in the upgrade it won't be called. It's just taking up a little space, nothing more. Look at old versions of WordPress upgraded over time and you'll find some redundant columns in some of the tables there no doubt.

If the CC developers could inform you of the DB changes/additions, perhaps in a simple list of SQL commands, then you could upgrade and then double check the DB by re-running the SQL commands in phpMyAdmin. Alternatively contact this person and ask them for their output, reduce it down to important issues and check/add those in after the upgrade. But I'd just see if the CC developers could possibly help you first.

At the very least, you just need to know any new columns/tables added.

And of course, back up first

[familychoice]

Most of those inconsistencies are non issues. The auto increments are of no real concern providing the column max value can cope, which it should be able to go to millions fine. The column orders are of no issue at all providing the mysql code has been written properly (and I would imagine, well hope, it has!), the primary key isn't a major problem really, again providing the code is written properly, you don't have to have a primary key, it's just usually set to enforce rules for protection. And the full text setting is again, just an order thing, but has no bearing.

Thanks for that Sarah, it sounded pretty serious but if it's not going to affect the store then it might still be worth a go. Someone's posted an update to the upgrade issues here: http://forums.cubecart.com/topic/44333-so-many-upgrade-problems-heres-why/ and it sounds like they're working through the issues.

The SEO URL algorithm has apparently changed though resulting in changed urls for products, so I'll need to check with the client to see how he's been promoting his products and if this is going to be a big issue.

Think I'll install a CCV4 demo (if they still have one available), import my clients database, then install a demo of V5 to see how the site fares before making any promises.

[suedenem]

Think I'll install a CCV4 demo (if they still have one available), import my clients database, then install a demo of V5 to see how the site fares before making any promises.

Surely you'd do any sort of upgrade in a staging environment, using a recent backup of data, and then thoroughly test before rolling out to the live site?

It never occurred to me that this might be an afterthought for some people - it's something I'd presume to be standard practice (even with incremental WP updates).  You tend to have a lot of problems with suppliers.  Perhaps it's worth looking at your processes to see if problems can be mitigated somewhat?

[familychoice]

Sue, sometimes you give great advice, other times....

Yes, of course I would test this out beforehand. Don't start the high horse stuff again please, I was just starting to like you.

[suedenem]

That's what I was afraid of.

[familychoice]

Think I'll install a CCV4 demo (if they still have one available), import my clients database, then install a demo of V5 to see how the site fares before making any promises.

Surely you'd do any sort of upgrade in a staging environment, using a recent backup of data, and then thoroughly test before rolling out to the live site?

It never occurred to me that this might be an afterthought for some people - it's something I'd presume to be standard practice (even with incremental WP updates).  

If you read my post properly you'll notice the demo test is to check for issues before I even quote for the job and recommend it as an option. *If* there aren't too many issues then I'll know how much to charge them for the job, and if it makes a mess of the URL's how much it'll cost them for me to clean it up. If it's full of bugs and doesn't pick up all the product settings then I'll recommend they upgrade to the latest V4 instead until they sort out the update script.

You tend to have a lot of problems with suppliers.  Perhaps it's worth looking at your processes to see if problems can be mitigated somewhat?

Sue, I know I come across as a bit of a numpty sometimes but I've been building websites for 20 years and running a successful business for the last 10. If the last few years have taught me anything it's that web software companies are more than happy to release bug-ridden buckets of poo, take your money, and then remove all means of support when said poo hits the fan. I'm not such a complete flippin idiot that I'd just slam in the upgrade without checking for issues on a test server first, like some of the poor buggers I've read about. In fact I'm not even going to recommend it until I've tested it out and checked out the experiences of other users - hence this post amongst others. The last thing I need is a repeat performance of the Trading Eye debacle.

That's what I was afraid of.

I still like you, but the wedding's off.

[suedenem]

Furry muff.  I read the bit I quoted and took 'promises' for 'changes' (in the middle of a big project that has to be out by month end, with DDN being a distraction).

Don't forget to charge for the research 

[sarahA]

Don't forget to charge for the research 

OOooo can I invoice for my response?

Oh just give me more cake when we're next passing

[familychoice]

Don't forget to charge for the research 

I might get away with including it in the job if it's a go-er, but I'm usually reluctant to charge clients if the results are negative. I know that's wrong so you don't need to tell me On the plus side, if it does go well I've got another ten Cubecart sites that might benefit so it's roundabouts and slides.

OOooo can I invoice for my response?

Oh just give me more cake when we're next passing

Is it still one slice of Victoria sponge per hour for this sort of work or are we looking at chocolate cake rates?

[sarahA]

Victoria sponge, it wasn't really of chocolate cake calibre answer

[suedenem]

I might get away with including it in the job if it's a go-er, but I'm usually reluctant to charge clients if the results are negative. I know that's wrong so you don't need to tell me On the plus side, if it does go well I've got another ten Cubecart sites that might benefit so it's roundabouts and slides.

I wouldn't say it's wrong, as such.  We'd usually charge for the investigative work if the client requests it.  We probably wouldn't if we initiated it as part of approaching the client for work (like your 10 CC sites).  All things to weigh up...

I don't buy into the not-charging-if-results-are-negative mindset, though.  A properly-executed test with negative results is a successful test.  My reasoning, should the client query the invoice, would be that i) we did the work and ii) that hour spent probably saved £xxx,xxx with a potentially botched upgrade.

Incidentally, if you do work that you don't charge for, still add it as a line on invoices (with £0 next to it).  That way, the client will see that there was actually work involved, reinforcing the value of your offering.

OOooo can I invoice for my response?

One of our clients forgot the URL of her analytics suite.  I emailed her the details earlier this morning (I knew the URL off the top of my head), and she'll be invoiced £25 for it

[Steve Lampkins]

Google slip me a couple of quid every time I post this link: http://goo.gl/Tvzvw

[familychoice]

Victoria sponge, it wasn't really of chocolate cake calibre answer

That's great value, I have to pay Sue two slices of chocolate gateaux and a mini-banoffee pie just for insults.

I don't buy into the not-charging-if-results-are-negative mindset, though.  A properly-executed test with negative results is a successful test.  My reasoning, should the client query the invoice, would be that i) we did the work and ii) that hour spent probably saved £xxx,xxx with a potentially botched upgrade.

Oh yes, I agree definitely. But in this case part of me thinks maybe I should already be up to scratch with installing the latest upgrades and already know any issues that might arise, so I'll probably file this under 'staff training' instead.

I want to give it a spin anyway, as even if it is a pain moving from 4 to 5, I'll need to know if it's any good for future jobs, though to be honest I'm very happy with the Open Carts trial at the moment.