I know, people are going to claim that a part time support contractor for VB can't be impartial. I can only say that I disagree. Over the years I have not been a fan of VB5 and those who have come to me for personal recommendations can attest, I could not not in good conscious recommend sites upgrade to VB5. For the first couple years I didn't recommend anyone use it. For the last year or so I started seeing how vBCloud has become a product that fits the segment of the market of people who want to have forums, quickly, but not be responsible for installing, maintaining, or otherwise keeping the site updated. Those with no knowledge of PHP or MySQL, and no intention of learning. However, I still was of the mind that anyone with large, established, forum stay away. I didn't think it was quite ready, in fact, part of me believed it never could be. I am surprised to say, I see now, I was wrong. Recently, in order to help another tech solve a possible issue related to impex imported content after an upgrade to 5.3.0 I decided to upgrade the one site I had access to that had Impex imported data- my main off topic discussion site with some 3.5 million posts. Now, I didn't bother with attachments but I have very few attachments for a site that size, I have always preferred members link to images rather than attach them- most usergroups never had permission to upload them. But everything else, avatars, post icons, smilies, ranks, notices were used. Also this VB 3.8 site is running on a shared plan on URLJet. I was going to install the VB5 version in a different (but still shared) URLJet account I have so I made the backup of the database SQL file, downloaded it, uploaded it to the new location, and asked the host to please import the 2.5 GB .sql file into a new database. I then uploaded the 5.3.0 files, setup the config.php file, and began the upgrade script, not truly confident it would complete. Now I knew from experience the upgrade would take hours, that is just the nature of upgrading from 3.8.x to 5.3.0 with over 3 million posts and an unknown about of PMs, all of them becoming "nodes" in the radically different VB5 database structure. I started in the afternoon and knew it would run through the night, if it didn't get a fatal error. It had run for at least an hour when I stopped checking on it even couple minutes. At some point it hit some seemingly random error, whether it was stopped for 1 minute or 2 hours, I didn't know, but I refreshed the upgrade.php page and the upgrade continued from where it left off. Bed time came and it was still chugging away. I kept the laptop open to the upgrade screen beside my bed as I expected it would require a refresh again. (I wondered if there was a way to code in an auto refresh if the upgrade is stalled for too long...) but anyway sometime around 1AM I woke up, saw something about a timeout, and hit refresh... again around 3am I hit refresh. Around 5am I saw it was stopped but didn't read the text, just hit refresh. Finally when I was ready to get up for the day closer to 6am I looked again and was ready to hit refresh but in fact it was the upgrade complete screen. I realized it may have been the same screen at 5am, I didn't look close enough- in which case I re-ran the final steps of the import an extra time- no harm done. So the upgrade completed, I deleted (renamed) the install folder and was eager to log into the Admin CP. Except I couldn't because the forum was closed. Wait.... what? Oh yes, the bug introduced in 5.3.0 (and fixed in 5.3.1) if you close the forum before upgrading (as you generally should) and you go to sleep (like me) so that your Admin account logs out due to inactivity, you can't log back in without using the tools.php file to turn the forum back on first. Oh well, but easy enough. So the forum is back on. The admin cp is familiar. The homepage (surprisingly) loaded fine. No avatars yet, forgot to copy them over, but that was easy enough, as were the smilies, as were the rank images, as were the default avatars (besides the custom avatars I copied first....) That looked better- it looked like a forum. Then I clicked on a forum with 1.5 million+ posts in it. And... Some long error. Exactly what I suspected! I affirmed my initial opinion of the software. But... I did know that VB5 still has issues when there are large numbers of forums, usergroups, and I don't know, maybe styles. And let me tell you, since I started my forum in 2008, I never deleted a single one of those things, though I only actually use a small fraction of them. So I visited the Admin CP. First I deleted 15 VB3 styles, all of them incompatible with VB5 anyway. (Wondered if there should be an option to delete old styles during upgrades?). Next was a look at forums (channels in VB5 lingo.) Sorry to say I had like 40 unused forums over years, I got rid of them. Finally, usergroups... Probably another 10 unused and unnecessary usergroups gone. The ironic part is that beginning in 5.3.1 none of these deletions would be necessary as it will finally stop loading everything into memory on every page load. So... back to the front end. I tried again... sure enough, it worked. I tried an even larger 1.5 million post forum, it worked too. The speed wasn't that bad. Then I remembered I could upgrade to PHP 7.0. (7.1 wasn't available on the shared server.) Once I upgraded to PHP 7.0 I realized I was browsing posts and forums with similar speeds to my VB 3.8 forum still running PHP 5.6.x. Further, things were working. I uploaded a logo, changed to a different built in theme, and it looked, and more importantly ran, like a forum. And that is to say nothing for the Blogs, Articles, and Groups so easily accessible now. (Yes I forgot my 3.x site actually had groups. They were almost never used, but there they were.) It wasn't a seamless upgrade, there were hiccups, and it would not have been as easy for someone inexperienced to upgrade as I did. I wondered what it would be like for someone who isn't me to try to upgrade their 3.8 site to 5.x while needing to make a support ticket for each issue found and waiting for a response, even on the assumption they explained the issue enough to get the correct answer on the first reply. It certainly underscored the need to upgrade a test site first so you (as a customer) have a chance to see what you may need to do during or immediately after an upgrade. Still, here was VB5, working, on a 3.5 million post forum, and working well. I had never considered upgrading my 3.8 forum to VB5 but for the first time I was actually considering it. It was a fundamental shift in my thinking. Prior to last week when I did this I honestly did not believe large forums could succeed on VB5. (Now I know, 3.5 million is nothing compared to forums with 30 million posts) but still, it is one, possibly two orders of magnitude higher than I ever previously considered. Certainly VB 5.x improved far more than I had given it credit for. Perhaps in my role as a support provider I was seeing nothing but problems reported, daily. People don't often make support tickets that say "All's well, no problems today." There had been the occasional posts on the vBulletin support forums from people happy with VB5, even defending it from harsh critics.I tended to stay out of those discussions. Realistically, will I upgrade my live site to VB5? Probably not. My members are still anti-change. The only members I still have are those angry other forums have moved away from VB3 to VB4, Xenforo, and other software. As this upgrade was happening I got a sincere thank you from one of my members for not upgrading the forum- unsolicited. They just wanted me to know the fact we were still on VB3 was something they appreciated. This person probably doesn't know it's a VB3 forum, he just knew that 8 years ago almost all forums were what I'm still running, and over the years, many of them have changed to something else, and he's disliked all the changes. I suppose my forum will be one of the last 3.x holdouts but for the first time I can't claim a non-viable upgrade path within vBulletin, it's because my users hate change. And that is a tremendous change in itself.