This site no longer opens on the first attempt.

[TrashyMG]

Not sure about problems, I haven't noticed any... however I have notice the amount of spammers creating fake accounts have dropped significantly lately. I was sometimes seeing 4-5 accounts being created a day that were obvious spammers.

 

[netcat]

I ranted about it on communication cube: https://pyra-handheld.com/boards/threads/the-communication-cube.59464/post-1706388

 

[netcat]

Because we don't want to have any security holes in there.

Software providers say they are patching security vulnerabilities to scare us into upgrading but the real reason is they want everyone on their latest version because it's much less expensive for them to support & interoperate with. The most secure version will be the old one that's been patched without functional changes, but they don't want you to know this. Trust me - I am a software provider. I have felt for decades that upgrades (software & os) are generally absolutely not worth the effort of getting everything back to the way it was before. Unless you really want a new feature, which is rarely the case these days. Android gets screwed up with each major update they force upon me. Basic stuff like connecting spotify in the car requires effort to get back (and causes me to panic - do i have to get a new car?!). Even the genius googlers with infinite resources can't get it right. Running apt-get dist-upgrade has never been worth it. I prefer to hack it so i don't get the "upgrade now" messages. It's a much smother experience overall, less trouble shooting. If it ain't broke, don't fix it.

If you are worried about the site getting hacked I would first ensure it's backed up adequately (so you can recover & have no data loss) then setup intrusion detection. These things are more important than ensuring your webserver runs in isolation, using a proxy, reviewing @ClockworkCoder's code or even pen testing.

The reason I mention this is because the site has def become more glitchy recently and PHP upgrades would explain it. And this is one of the two internet sites I choose to use. I like it.

So: software providers are outright lying to us. New features = new bugs. Always. I suspect one day the EU will regulate forced upgrades.... sad but true.

 

[netcat]

Also..... if you are security conscious it's odd that you have so many ports open:

nmap -Pn -T4 -p 1-65535 -O pyra-handheld.com
Starting Nmap 7.91 ( https://nmap.org ) at 2022-05-21 13:38 BST
Nmap scan report for pyra-handheld.com (65.21.92.44)
Host is up (0.046s latency).
Other addresses for pyra-handheld.com (not scanned): 2a01:4f9:3b:4453::10
rDNS record for 65.21.92.44: amaterasu.dragonbox.de
Not shown: 839 closed ports, 142 filtered ports
PORT STATE SERVICE
21/tcp open ftp
22/tcp open ssh
25/tcp open smtp
53/tcp open domain
80/tcp open http
110/tcp open pop3
143/tcp open imap
443/tcp open https
465/tcp open smtps
587/tcp open submission
993/tcp open imaps
995/tcp open pop3s
1717/tcp open fj-hdnet
1718/tcp open h323gatedisc
3128/tcp open squid-http
4000/tcp open remoteanything
9418/tcp open git
10010/tcp open rxapi
20000/tcp open dnp
Aggressive OS guesses: HP P2000 G3 NAS device (91%), Oracle VM Server 3.4.2 (Linux 4.1) (90%), Linux 2.6.32 (90%),
Linux 2.6.32 - 3.1 (90%), Linux 5.0 (90%), Linux 5.1 (90%), Ubiquiti AirOS 5.5.9 (90%), Linux 5.0 - 5.4 (89%), Ub
iquiti Pico Station WAP (AirOS 5.2.6) (89%), Linux 2.6.32 - 3.13 (89%)
No exact OS matches for host (test conditions non-ideal).

Also odd that http headers indicate you are using apache which afaik is know to be less secure than other webservers.

Sorry to be such a jerk but I think PHP upgrades are a misguided security tactic.

 

[EvilDragon]

It wasn't an upgrade to a new major version, I only do security updates until the software I use (in this case Xenforo) gets a new major version as well, as in this case I need to retest everything anyways.

No idea why, but for some reason the security update disabled the cache which was the reason the site was overloaded and lead to some issues. I don't really see a reason it did that... maybe a mistake someone made in the package, but it happened.

Almost all open ports you listed there (except for 2 or 3) are being used and therefore need to be open. This is a dedicated server that's used for web, email, git, ftp, etc.

Every single service is protected by fail2ban and similar and doesn't spit out detailed information that can could be used to figure out what security holes the server should have.

I regularly follow newly found exploits and check whether they apply to the server and software and close them.

I have regular backups (daily), so I'm not concerned of someone breaking in and deleting data. But I have tons of customer details in the shop, so I don't want anyone to break in and steal the data.

That's the reason I install all needed security fixes.
I doubt it's a good idea to not install security updates simply because the software is running good as is and ignore the fact that a lot of newly known exploits will find their ways into the next generation of bots trying to steal as much data as possible.

 

[netcat]

Ah ok. Sensible.

The issues I have seem to be related to repeat visits:
- Previously viewed thread opened at first post (if no unread posts)
- ... or at very bottom of last page beyond posts (if I do have unread posts)
- "Jump to new" buttons not always available

I'll bet if you asked a PHP developer they'd say it's the fault of a dependency - that they were coerced into upgrading.

 

[levi]

I had a look at the issue with the page first rendering somewhere south of the actual last unread post. It looked to me, looking at the source that the anchor tag was in the right place, as part of the post opening framework. So I think it's a matter of your browser being less exact than expected, rather than the site producing bad data.

I don't think this is the exact right thread to report this to, but it has come up again so this'll do.

 

[netcat]

Yep - it's the software blame game.

 

[TeDaDeS]

I've not spotted the half-loaded page anymore. So that might have been the 'cache not being enabled' issue.

 

[ClockworkCoder]

Yes, it's working much better for me too now.