seawolf

So, why do you think they will have unworkable failure rates? The cooling system is ingenious and works very well from all reports. They are also made in the USA probably to some of the best quality control standards money can buy.

I would stay away from it and look at Ruckus or Ubiquity Unifi instead.

I suppose it depends on what exactly you mean by "real" enterprise class hardware. Here's what I suggest, the next time a vendor says that, ask them to explain what they mean and provide some specific examples. If their explanation sounds like b.s. and/or you would require a degree in electrical engineering to understand it - they probably don't know what they are talking about. If they can give you specifics and areas of differentiation in a way that a human can understand it they might have a point. The Ubiquity Unifi solution has a a software controller that can be installed on a server (Linux VM recommended), so it's not really a controller-less system. Look at it this way, they are inexpensive enough to run some real world test in your environment without breaking the bank, and then you can see for yourself. Also, Ubiquity does provide support.

I've tried a couple of the Cisco Small Business switches (Linksys) about three years ago, but had one of them blow up on me after 6 months of use and with the other one, LACP stopped working after a couple of years just out of the blue. Real Cisco switches are great - expensive, but great. I stay away from the Small Business line now, but maybe they've improved. Anyone with experience with these lately?

Personally, I find Meru to be more about talk and marketing than great technology and solid WiFi performance. Have a look at some of the very in-depth, real world conditions testing that Tom's Hardware did a few years ago with some of the most widely known and used WiFi system. Ruckus came out on top, and Meru wasn't tested. Why Your Wi-Fi Sucks And How It Can Be Helped, Part 2 - Welcome To The Wi-Fi Cage Match Meru have always claimed their single channel architecture was superior way before the advent of 802.11ac. However, the problem with the Meru architecture is that all the APs on a "blanket" share the same channel. So, coping with interference is extremely difficult on that "blanket". So, despite Meru's claims, they are going to run into the same problems with 802.11ac although they may achieve higher speeds using the architecture than other APs, they will do so with higher risk of interference as they have had in the past. I have not tested the new Meru APs though, so I could be wrong. But, If I were to build a WiFi network all over again, I would either be looking at Ruckus or Ubiquiti.

I'm not surprised your sole user has had issues. Here are some things you are likely to come across at one time or another: 1. Inability to bind to AD domain or losing the AD domain bind. This seems occur periodically with various releases on both the Mac and Windows server side 2. VERY, VERY, VERY slow network account logins, sometimes exceeding 5 minutes on the LAN, and taking over 20 minutes for logging onto mobile (network) accounts off the LAN due to the number of timeouts that occur. 3. VERY slow mounting of network drives, slow copying to said drives. 4. VERY slow printing to network printers 5. Major bonjour issues, including for AirPrint and AirPlay on iOS devices. Some links with discussions about these sort of issues commonly faced in .local domains. As I said, .local domains were a cruel joke played on the world by being in Microsoft's "Best Practice" documentation that was then espoused by MS technicians the world over. https://discussions.apple.com/message/15834652#15834652 OS X 10.7, Lion, and ".local" domains | Thursby Software http://www.centrify.com/downloads/public/centrify-directcontrol-for-mac-local-domain-workaround.pdf Dmitry Dulepov: OS X Lion and local DNS issues active directory - Painfully slow login to AD bound Mac OS X Leopard machine when off home network - Server Fault https://jamfnation.jamfsoftware.com/discussion.html?id=8872 https://jamfnation.jamfsoftware.com/discussion.html?id=6581 https://discussions.apple.com/thread/2143988?start=105&tstart=0 https://discussions.apple.com/message/12346886#12346886

The iBoss web filter by Phantom (http://www.iBoss.com) is capable of intelligent image blocking for Google and other sites. It is also the best web filter I have ever seen in all respects, including reporting.

We've been using Ruckus for 3 1/2 years now. Works great and the new APs are especially good.

One Mac is one thing - 300+ on a .local network is another all together.

We have two virtual DCs and one physical DC. The physical DC is recommended in a virtual environment as so many systems are dependent on DNS these days and it also means people can still authenticate if you have issues with your virtual environment. We use a low-spec Dell R210 for the physical DC. It has 4GB RAM, but not much grunt overall. We backup our VMs using Veeam to a FreeNAS server and then have Veeam setup to do a secondary copy to a Drobo B800i connected to the backup server via iSCSI. It is recommended that you always keep at least two copies of your backups. And that you verify them occasionally.

+1 for BackupAssist. Low cost, flexible, reliable, and easy for even noobies to understand. Veeam is a better backup solution for Hyper-V or vSphere. It is also a lot more expensive and fickle.

Agreed, never run a DC (or anything else really) on a server you also use as a Hyper-V host.

Cause a "few" problems - that's an understatement! I would rather have my fingernails pulled out one by one than try to use Macs on a .local network. It can also cause problems for other systems as well.

I agree with the never using a .local domain, that was a cruel joke MS consultants played on the world. Also the basic structure SYNACK stated is also a good start. However, I disagree with the advice about not using a valid domain name - what is the reason for that? It's actually considered best practice, and I have never encountered a single problem with it. You do need to either register a domain name that will never be used publicly for the internal domain or use a subdomain of a public domain and use split DNS (views) to properly segregate public from private. I could only see a problem if something were misconfigured quite badly.

Well, take heart. What you saw in the 2010 pics cost $47,000 all up including a day of work by an electrician, 4 days work for the data cabling (200 port frame to extend cables was required), the racks, two UPS, 3 servers, a SAN, a firewall, two switches, KVM, and all of the software licensing. No, 47k isn't a small chunk of change, but it took quite a bit of haggling to make it stretch far enough to do everything you see. The rack that you saw in the 2010 and today pics is the same rack, just with extra add-ones and modules I purchased for it. I like to buy gear that can expand and have a long lifespan. As for the Suns, we currently have one x4150, two x4170s (M1), a 7110 (4.2TB), two X3-2's, and a 7320 with a DE2-24P disk shelf. The X3-2s and 7320 were purchased in the past 4 months and serve as our core vSphere infrastructure. The others were purchased in Dec 2009. I'm using one of the older x4170 as a Hyper-V server and one as a backup server (Veeam and BackupAssist). The x4150 is used for testing and the 7110 is a backup emergency SAN as well as serving as storage for us in the IT Dept. I try to keep good equipment for as long as possible and have in the past had a 14 year old Sun Netra server that was still running as reliable as a Swiss watch.

I would fire the company that did that without a seconds thought.

This does sound a lot like STP and port blocking. Even if the switch says it is disabled, it may actually be taking effect. I've had firewalls with buggy firmware telling me traffic was allowed in the logs when it was clear and evident it was actually being denied. Could be something similar going on here especially if the HP support guys say the logs look fine. If you do end up having to replace these 1810s and only need unmanaged switches, the Netgear unmanaged switches are the very best bang for the buck you can get. If you need managed switches, the HPs are the best value (despite your issues with these 1810s don't let it burn you on all HPs).

Well, I certainly can't fault your attitude or courage. However, wisdom is not really a matter of opinion, but based on hard earned experience, and learning from that experience and the advice of others with greater experience and knowledge. If you think that being on the bleeding edge is always going to end in sunshine with 100s and 1000s sprinkled on top, you are in for a rude awakening. Continue on that path and you will suffer the wrath of users and your sanity in the foreseeable future. Count yourself lucky that things appear to have gone so well this time around, but understand that it would be foolish to believe that string of luck will continue forever. Ignore this advice at your own peril.

The noise level in the server room is 79db standing at the racks. The equipment in the room consumes 480kWh per week or just over 2000kWh per month.

We did everything in house except for the cabling to the patch panels and the Lino flooring. For this last upgrade where we removed everything from the room and replaced the flooring, the system was down for 40 hours in total. We took it down on a Wednesday evening, pulled up the floor and painted until 2am. Floor was layed the following morning and we reinstalled everything the following morning. We had everything back in and spun up by the end of the day, although there was still a lot of temporary cabling between the servers and switches strung across the floor for a couple of days. The biggest delay was caused by having the flooring replaced, but it was time to get rid of the circa 1985 carpet once and for all. The original upgrade back in 2009/2010 consisted of replacing everything (nearly ever switch in 12 buildings, as well as all of the old Win 2000 servers that were running things at the time and creating a new domain and user accounts. That outage was much longer at around 18 days in total. I could only do something like that once or I would be strung up.

LOL. Well yes an average size person can fit in that space quite nicely. And the noise from the servers and switches provides nice cover as well.

Over the past 4 1/2 years we've put quite a bit of effort into getting our core infrastructure up to scratch. I've put together a little picture documentary (attached) showing how it has evolved since late 2009 to today. I'm sure there are some great setups out there in the community and I would love to see some of them. Or even if you just have what you wish was only a "Before" scenario and would like to get it to a "Now" condition and would like advice from people in the community who have been there and tackled many of the issues you will encounter (technical, financial, and with change management) and can assist with getting you to where you would like to be.

Don't you have a config backup that you could restore on the firewall to get it back to a working state? If not, then big lesson learned for you. Configuration management begins with backing up your configs. It is clear from the logs that it is the firewall rule "OUTSIDE_access_out" that is causing the problem. I can't see a firewall rule with this name in your screenshots though. I do note that you have several "Ruckus-Squid" rules disabled. Are those deprecated rules or should they be active? p.s. I wouldn't recommend posting firewall screenshots showing all of your firewall rules including ports and IP addresses on the open net.

There used to be quite a few case studies on the Ruckus website.

Is that a home grown solution (very impressive if it is), or a commercial solution? If it's commercially available, who sells it? Cost?