mrbios

Assuming no databases are running on the 2012 instance you can safety remove Database engine services and SQL server replication and just leave the rest which won't take up much space anyway.

Hmm, well this morning everything went fine. CPU usage on the server didn't go over about 25%, 30mins later the wonde service did a sync and that knocked it up to 40% with some expensive queries but that didn't last long. Maybe it was as simple as the compatibility setting...

Where would those errors normally present themselves? I'm not seeing any errors in event logs, sql logs etc. I've got a report i can run as a test that seems to take anywhere from 47 seconds to 90 seconds to run which pulls attendance data for a specific year, but nothing i change seems to make any notable difference to the performance of that report running...so what I'm really struggling with is actually testing the problem, as i don't seem to have a constant i can refer to. And this particular performance issue only seems to be presenting itself in the form of edulink attempting to do things when everyone uses it for registration in the morning via some tablets. No one has complained about sims being slow or not being able to achieve something in sims either, so the issue seems isolated to that heavy use case with morning registration.

It is now, it wasn't previously. However from my testing i don't really see any difference. Is it likely to make a difference to an individual - such as the amount of time it takes to run X report, or is it likely to make a difference in terms of scale, multiple users performing actions for example? The issue I'm seeing appears to pop up when lots of registration queries and writes are going on at once. I can't seem to cause the issue when doing one big report or command myself. Is there any solid way to test this perhaps?

I recently migrated our MIS/SQL server from Windows Server 2012R2 and SQL Server 2012 to Windows Server 2019 and SQL Server 2016 and i seem to be hitting some CPU related performance issues, specifically in regards to using Edulink. The SIMS database is roughly 20GB. Total size of all databases on the server is about 28GB. The SQL server has a maximum assigned memory of 28GB, and the VM is set to reserve its entire memory allocation. When on 2012 the server had 32GB ram and 4vCPU, now it has 36GB ram and 8vCPUs. I've set MAXDOP to 4, and cost for parallelism to 20 as a slight increase over the default of 5 (I've tested quite a few settings on these two in an attempt to resolve this, but to no avail) In the morning during registration when we have 30-50 devices all using edulink for registers at the same time the CPU usage on the server will spike up to 70-90% and the edulink integrator app will start reporting an issue that requests are taking too long to process, and people trying to open a register will see a timeout after 45seconds. Anyone have any suggestions? It makes no sense as so many poorly configured aspects of the old server are fine tuned on the new one, and i can't understand what could be causing this.

"get the spray paint out, draw a circle on it, deal with it in about 6 months. Sorted" - The council, probably.

Bin fell over. It'll be ok, we will rebuild.

Intel NUC with Batocera on it. Fill it with retro games That's what I've done. I've got about 7000 games from nes, snes, n64, ps1, atari, amiga, old arcade mame games, gameboy. Then connect an xbox controller over bluetooth to it, or buy some usb retro controllers like SNES ones. Instant fun... of course obtaining the roms of these games isn't strictly legal unless you own the physical media, so I'm obliged to say i own the media..... I've got a decent nuc that can even run WiiU games like donkey kong tropical freeze lol.... although, it's not as smooth as playing it on a wiiu.

Long story short, I'm trying to convert amazon.com $ gift vouchers into £ and buying one of these from Amazon US with the gift vouchers and selling it on seemed like the easiest way (there was probably an easier way...) £260 posted, any takers?

After some advice from the hive mind of edugeek... I've got one spam filtering rule for the whole school, the same settings apply to me as applies to the member of staff having an issue. The county council are trying to email a member of staff but every email ends up being filtered as spam by our spam filtering policy. If they CC me into the email, it comes through to me without any issue at all. I've got gloucestershire.gov.uk (the county council) as an allowed domain in my spam filtering policy, so they shouldn't be getting filtered by it at all. There's nothing i can decipher within the message header to indicate an issue. I'm at a loss as to why this is being blocked, the email literally contains text, no attachments, and one link to the gcc website... so there's no content it could be picking up on either. Any ideas where i go from here to determine what's going on?

I mined crypto back in the early days of btc-e and mtgox, bought two AMD graphics cards at the time, mined the value of the cards, then gave up and was super happy with £500 worth of "free" graphics cards lol. Reinvested in eth when it was £250 a coin, doubled money again and then cashed out. (Obviously regret that now lol) Invested properly again earlier in the year, now hodling long term on eth and mdxt. mdxt is a relatively little known coin with a genuine purpose, low market cap, only $0.10 a coin at the moment, but could easily be similar to FTT in terms of coin value the future given it's an exchange utility token that works in a similar way to BNB (or will eventually). Projects still fairly young, waiting for some aspects of the mandala exchange to finish development. I've got some friends that are heavily into NFT's who are holding cryptopunks and rarepepes, i'll never understand the NFT craze however much they explain it though, i'll stick to good old coins EDIT: Tips: never tell anyone how much you hold. Don't encourage family/friends into it, or if you do, make them well aware of the risks. Always double check addresses, and do test transfers if sending large amounts. Staking can earn decent returns. MFA everything!

I changed my offices direct dial number last year, now don't give it to any suppliers, and the receptionist screens my calls. If it's a cold caller she takes down the number and i block it on our PBX. This has significantly reduced the amount of time i've spent on the phone. (If you don't have a willing receptionist, try palming that role off to an unwilling tech ) Millgate have been ace for me, tend to get really good prices and have always been a pleasure to deal with. Would probably put them down as my #1 recommendation for quotes. EDIT: As LeMarchand just said, it'd be an insult not to mention ICT Direct, they're great and always worth keeping on your most used suppliers list!

I've never been a fan. When GDPR came in i questioned it with the DPO but he was happy. Only consideration the school have made about changing it over the last few years was down to how annoying it was in teams, but then we changed their display names from username to full name and they forgot about that too. I might try and change it next year, though i can only see that being a future change rather than rolling it back to all existing pupils.

Students follow: 10JBL2512 Year of birth First letter first name First two letters surname Date of birth Month of birth It's quite irritating, but it's been that way for ~20 years, so I'm reluctant to change it Staff are just firstnamelastname.

The support guy ruled out routing, he did do a bit of testing on that, but everything seemed fine from a routing perspective.

So after some diagnostics with Smoothwall support, and looking at more in depth logs, it looks as though the fault with this isn't with smoothwall. Smoothwall is sending the accounting packets, NPS is receiving them, but isn't sending an acknowledgment back. Not that surprised really. Might try setting up either a Server 2022 or even a 2012R2 server just to test and see if it's a 2019 specific issue (as i know it did have some strange NPS issues in the early days)

Interesting, I'm still on 48, got a smoothwall support call tomorrow morning about this. Maybe i should update to 49 overnight and see what happens

I'm trying to move a trust-e install to a new Pc at the moment but im getting stuck at a password screen. I've followed everything in this thread, our reval unit works and i think trust-e is all installed etc however when i run trust-e.exe it asks for a password, but a different password screen to the one you get when you login to trust-e. I've tried the usual passwords that log me in to the trust-e software, then i get an error of the following Error Number: -2147467259 Error Description: [DBNETLIB][ConnectionOpen(Connect()).]SQL Server does not exist or access denied. In module: modJG_ADODatabase In Routine: OpenADOConnection If i click yes to retry error description changes to: Invalid authorization specification Anyone come across that one? EDIT: Scratch that, looks like i copied the wrong NRSCashless ini file.

https://h10145.www1.hpe.com/downloads/SoftwareReleases.aspx?ProductNumber=J8699A That what you're after? Looks like the bootrom version you need are near the bottom. Handy tip, replace the J8699A with any HP model number to find the firmware for any HP device.

I restarted my smoothwall box last night. Interestingly, wireshark is still seeing the one occasional packet received, but from what i suspect was the very first device to connect and report via radius. I think smoothwall struggles to report back properly using multiple radius clients rather than a single IP address from a controller. EDIT: I've just gone through and added all 49 access points individually, rather than the range they're in. Now ideally i want to restart the radius server service, but i've no idea how i do that without restarting the whole smoothwall? EDIT2: Nevermind, worked that one out, restarted the freeradius service but no change.

Just tried that but doesn't appear to have made a difference unfortunately, enabled firewall logs as well and i've got the following 2021-07-12 16:54:37 ALLOW UDP 172.18.66.100 172.18.66.39 1814 1813 0 - - - - - - - RECEIVE 66.100 is my smoothwall, 66.39 is my NPS server. So I'm definitely getting something through, albeit wireshark shows that 4 times while the firewall log only sees it once throughout the period i had logs enabled. The more i investigate the more i feel like smoothwall is at fault.

That sounds exactly how i've got mine setup with the only change being that rather than an individual controller IP, I've got the full CIDR set as the authorised radius client range on the smoothwall. I've just done a complete config update and put the latest firmware on a bunch of APs and I'm still only seeing radius accounting packets coming through to the NPS server from smoothwall on occasion and always relating to the same individual AP, not all the others. I'll try adding in a bunch of APs individually as authorised radius clients on the smoothwall rather than the range and see how that goes. EDIT: Interesting, every time i add an AP, i get an accounting packet, but all are identified as that same AP I added first.

Yea the APs are setup in their own vlan so on RAIDUS the full cidr of their management vlan is setup as the client. Authentication is working fine, and i can see accounting is being received by my smoothwall, but forwarding that accounting back from the smoothwall to the NPS server doesn't appear to work.

I'm trying to get radius working nicely with smoothwall but i keep getting errors on the functionality tests. I've set it up so that my NPS server is doing authentication and the smoothwall is doing accounting, and then should be forwarding accounting back to the NPS server. However under the smoothwall functionality tests i keep getting the error "The system attempted to forward the RADIUS request to the upstream server, but it did not respond or the configured shared secret is incorrect." I'm certain it's not the shared secret, i've redone that twice. I'm running wireshark on the NPS server and i can see the occasional accounting packet going from my smoothwall to my NPS server, but they only ever appear to come from a singular access point (Under NAS-IP-Address attribute, the source is the smoothwall as expected) The other errors I'm seeing in the functionality tests: No Framed-IP-Address attribute was received with an accounting request, but DHCP is not installed. This configuration is not supported. Calling-Station-Id attribute is missing from the RADIUS packet. This configuration is not supported. NAS-IP-Address attribute is missing from the RADIUS packet. Wireless roaming may cause problems with unsynchronized logins and blocked users. Any suggestions? I don't really know what I'm doing on the Aerohive side of things so maybe i've not done something right there. It's actually ExtremeIQ or whatever Aerohive is rebranded to now.

I recently replaced my storage and servers (3 host, iscsi switch, san scenario) with a hyperconverged setup from StarWind. £35k in all including 5 years of proactive support. Previously had a Nimble san which was about £20k plus £12k for an expansion shelf 2 years later. Prior to that we had a super cheapo SAN....a typical <£10k box like people are mentioning here, and i can safely say that you get what you pay for! The cheap things people mention in a few comments above are crap for a network of any decent size. SANs are expensive and outdated though considering you can couple your hosts in with them for the same price as a SAN alone. You can go much cheaper again if you go down the route of storage spaces direct, which would have been my preference had it not been for my desire to stick with VMware over Hyper-V. Converged/Hyperconverged is the way forwards. However in terms of cloud, personally i still don't feel Microsoft have got it quite right yet with OneDrive. Last time i tried it with shared windows devices i ended up with too much of the: sign out too quickly and file doesn't get synced to the cloud prior to logoff problem. Great for any other scenario such as chromebooks, 1:1 scenarios etc though.