Dave Ambrose

It will be mailing out from the printet in a day or two.

Yes, there's a new magazine. Scott, the publisher, is a former staffer for MCM.

Interesting. I think I can explain your case. I was downloading a backup set from the server at the time you posted this. I've seen high network traffic slow things down, and will be more careful about when I retrieve those backup sets.

We can learn a lot from other modeling genres. My son and daughter build Gundams and other closely allied subjects. They've turned me on to all kinds of cool tools. We have a large Gundam model contest here in SoCal, and it's one of the nicer contests to attend. First off, I'm THE old guy in the room. But people are friendly and having fun with their builds. I've built a few, but I'm not big into the genre. There are lots of opportunities for cross fertilization that we should be using more than we are.

Maybe it's a nod to MCM? ? Galactic headquarters for the magazine are in Honolulu.

For me, it depends on the subject. Building a hot rod or dragster without an engine just doesn't look right. ? I once built a '60 Impala without the engine. I was very pressed for time, and it was needed for a museum display, so the engine wasn't necessary. Just glued the hood and trunk lid in place, then painted it. I've built a few curbside kits in recent years, and I have to say, they came out well. I don't miss the engine because the bodies are the primary element in the final build. That holds true for a lot of my models with engines too. If it means getting new and better kits sooner, then by all means, leave the engine out. We can either find one to put in there, or build it without. The increasing dominance of 3D printed aftermarket parts means just about anything can be made for a given kit and added for a reasonable cost. I think there's enough evidence to suggest that the majority of purchasers don't care about the engines unless they're very prominent.

I got to talk to him at a car show and have ordered parts from him. My experience ordering parts was good. They came a little sooner than predicted, and were very good quality.

Oddly enough, this isn't our first run-in with an AI project. Many years ago, there were a group of servers chewing up a huge amount of our bandwidth, and we were getting charged for the overage. The server addresses belonged to some AI project located in Fremont, CA. They never responded so I ended up blocking them.

I'm 90% certain that the first problem was someone using this site to train a large language model AI. I don't know about this second one, but it came from within our hosting company. I've already talked to them, and the problem is being referred to their senior level techs.

We had someone hammering our server again. Blocked them and it cleared the problem.

Joe, you should have gotten your replacement issue. Please PM me if it didn't arrive.

We had one or more entities trying to read out the site content. It effectively turned into a denial of service attack. This problem hit other forums too. I was able to develop an adequately selective filter to keep them from accessing the site. They may work around it, but it's sufficiently flexible that I can filter them out again if needed.

Because there's a bunch of white space in the post itself.

Made some modifications this afternoon, and they seem to be working. I also upgraded the forum software to the current rev. If you notice anything odd, please comment on this thread. Our sluggishness problem was related to a Chinese "search engine" continually accessing the site. I don't believe this is, in fact, a search engine. I believe they're training an AI model. They might work around our access blocks. Please comment here if the forum gets sluggish again. Thank you for your patience. This was a hard one to solve. -- Dave Ambrose

Update: Have confirmed that (ro)bot traffic is the root cause of our problem. Normally, our server runs about 5% load. These days, it's averaging 50%. I have some ideas for remediating this, and will be implementing them later this week. We'll have to see if this works.

Thank you! I hadn't noticed this, but I think it's an important piece of this puzzle.

Update Thank you all for being patient. In case it wasn't clear, we are still running on our old server. We won't be migrating until this problem gets diagnosed. This problem has been an exercise in ruling out possible causes. Current hypothesis is that this is being cause by a Huawei search engine crawling our site to index it. Normally, we would tell it not to make the troublesome accesses, but it's ignoring those directives and crawling the site anyway. When it does so, it loads down our server and causes a slowdown for legitimate users. Real search engines throttle their inquiries so this doesn't happen. Many other sites are reporting that Huawei's IP addresses have been hitting sites hard and causing performance problems. That looks like it might be the case here. I'll try to confirm that, then configure the server to block their access. It would help me if you can note the date, time, and your time zone when it's slow and post that information here. That will help narrow down my search. I may be taking the server offline at short notice as part of my diagnosis process. Again, thanks, I won't be quitting until this is solved.

Update Thank you all for being patient. This problem has been an exercise in ruling out possible causes. Current hypothesis is that this is being cause by a Huawei search engine crawling our site to index it. Normally, we would tell it not to make the troublesome accesses, but it's ignoring those directives and crawling the site anyway. Many other sites are reporting that Huawei's IP addresses have been hitting sites hard and causing performance problems. That looks like it might be the case here. I'll try to confirm that, then configure the server to block their access. It would help me if you can note the date, time, and your time zone when it's slow and post that information here. That will help narrow down my search. I may be taking the server offline at short notice as part of my diagnosis process. I'll be posting this elsewhere, but our server migration is on hold until this is resolved. Again, thanks, I won't be quitting until this is solved. @Ace-Garageguy, this wasn't done in anyone's spare time. There's actually a fairly large staff behind this forum software and they generally do a very good job.

Thanks, that's a very interesting clue.

I have mixed feelings about this article. First off, it comes from Microsoft, one of the active agents in creating the Crowdstrike debacle. That's a major conflict of interest. Secondly, they characterize the underlying problem as small. Ahem, catastrophic failure of kernel level code is never a small error. Third party analysis of the code indicates that the problem is triggered by a combination of poor coding and a possible data corruption. You need a special kind of engineer for this project. You need one that relies more on design, and less on testing to ensure correctness. It's a "do it right the first time" attitude that you don't see very often in any generation. They tend to look less productive than sloppier coders, so they tend to not get promoted, and leave for greener pastures.

So, was reading this afternoon that McAfee had a similar problem some years ago. Oddly enough the current CEO of Crowdstrike was CTO of McAfee at the time that all went down. Guess they didn't learn anything at McAfee. Also long rant about companies cutting personnel until they don't function very well.

I think a few people are going to get yelled at; both at Microsoft and Crowdstrike. Crowdstrike produced a defective update, but Microsoft was supposed to verify and authenticate the patch. I wouldn't fire them though. It will be hard to find anyone else with such a fine appreciation for the consequences of a mistake like this.

Look in your spam folder. Sometimes our emails end up there despite our best efforts. Your subscription starts with the next issue. As soon as we upload issue 225, it will appear in your account. If it doesn't, please let me know. -- Dave Ambrose

Problem seems to be a database issue. I'll be able to work on it this week.

I think Firefox is working well with the forum. This forum relies heavily on Javascript to make it work. If you have any enhanced security settings, that's likely to be the cause. Edit: I posted this reply and addendum with Firefox.