Something Random

Pradeep said:
When a tweeter is offset from the woofers axis, they are often designated as a left/right pair, and marked as such. Ideally you would have symmetry, such that at least you, in the prime seating position, would have the best imaging possible. If you have plenty of off axis seating positions, a speaker with a coaxial driver (tweeter sound comes out of the dustcap of a woofer) can help things out. The difference is usually noticed in terms of dialogue clarity, or lack thereof.

If you have two identical yet offset bookshelf speakers you might be able to flip one over to get symmetry.
Click to expand...

Thanks. I ended up buying a different kind of speaker (symmetrical type) as I decided not to take a a risk on the ones in question. Unfortunately I have to buy on the internet, so it is necessary to rely on user opinons and web photos.
 
ddrueding said:
Nearly two years old, but new to me.

A decent documentation of a 100TB DIY server.
Click to expand...
This chassis did not exist when the guy from your link built his server, but if I had to do something similar today, I would be more interested in this SuperMicro chassis. Smaller (4U), more power efficient (80Plus Gold power supply) and probably more future-proof since enterprise hard drives are moving more and more towards the 2.5in size. With 7 6500rpm 80mm screamers, it must not be any louder or much quieter than the Chenbro 9U monster though.

I don't know why no one builds a 4U chassis with 3 or 4 180mm slow-rpm fans. It would fit in and provide considerable airflow, without the noise assosciated with +5000rpm 80mm fans. If a single 180mm fan can cool a GeForce GTX580 inside a smallish Silverstone Temjin TJ08-E, it can certainly cool ten to twenty hard drives inside a 4U enclosure.
 
Make that 3 or 4 140mm fans instead of the 180mm. 180mm fans wouldn't fit inside a 4U chassis. Still, I'm sure a few 140mm fans would be quite enough for almost any configuration.
 
We just got a memory upgrade for our ESX farm. 60, 8GB DIMMs replacing 60, 4GB DIMMs. :beye:
 
ddrueding said:
Got plans for the 4GB DIMMs?
Click to expand...

Our internal lab team can re-purpose them for other uses. We went through planning of our hardware and found this was a better value for our ESX to extend their life rather than buy all new hardware this year. Our systems are mainly limited by RAM because the Nehalem-based Xeon's (X5550) are strong enough to handle more VMs.
 
My congratulations to Lenovo, for building a desktop that is simultaneously:

1. High-tech (No PS/2 ports)
2. Low-tech (VGA only)
3. High-tech (No Parallel)
4. Low-tech (Serial)
 
Last time I was at an actual computer store, I saw that somebody is making an ithing to RS232 adapter, presumably for console sessions on network hardware.

Also, remember that IBM (Lenovo) is a big player in POS systems, where RS232 is very much a living technology if only because there's a huge installed base of receipt and label printers out there.
 
Mercutio said:
Also, remember that IBM (Lenovo) is a big player in POS systems, where RS232 is very much a living technology if only because there's a huge installed base of receipt and label printers out there.
Click to expand...

Indeed. I'm swapping out all our serial/parallel POS crap to at least USB, network if I can manage it.

Mercutio said:
Last time I was at an actual computer store, I saw that somebody is making an ithing to RS232 adapter, presumably for console sessions on network hardware.
Click to expand...

That is an interesting thing. My only use for serial would be to remotely control my 4x4 HDMI matrix switch (IR and serial, gotta love the monoprice folks).
 
I thought I was getting good at my job. No significant loss of productivity company wide in the 3 years I've been in charge. No loss of data. No virus infections. And for the last 2+ years since I got most of it set up, no all-nighters.

I have 3 VMWare servers, any 2 of which have the horsepower to handle all the load. Theory goes that if a machine fails I can just shuffle the VMs and off we go.

One of the servers failed at 4:30PM yesterday.

Apparently one of the remaining servers, even though it has plenty of horsepower, goes unstable and becomes unresponsive under this higher load, even though it was fine with it's normal load. And the server that is stable regardless doesn't have the guts to handle all the excess load on it's own.

So here I am, 4:30AM, building a new ESX machine to deal with it. I need to get smarter before I get too much older, this is much harder than it used to be.

*this isn't a tech support question, just a rant. No need to try to help, I have it covered, thanks.
 
I think that's the point where you start actually paying for the ecosystem and getting the management tools. VMware for me is "I need to continue running a server system that was on old-ass hardware" not "I need to run dozens of single task guest systems with an iSCSI back-end and multiply-redundant everything."

But methinks that's where you are or at least where you're going.
 
I don't bother with the iSCSI backend, all the disks are local. The simple description of the current problem is just flaky hardware with the simple solution being better and more hardware.

Of course, if I wanted to implement systems and management capabilities that would prevent a flaky machine or two from becoming an all-nighter, VMWare has me covered. I just feel that the additional costs in licenses for the machines and training for me would get us very little at our current size. Perhaps the risk of an all-nighter every 3 years is the cost effective solution.

I'm thinking of just going with 2n redundancy instead of n+1 for the hardware and doubling down on backups via faster disks and a dedicated network.
 
I am constantly battling shitty USB to serial adapters in my job because the laptops that were approved have no real rs232 port.

Right now I have a dongle here with a prolific pl2101 chipset in it that works for modscan32 but not chipkin cas modbus scanner.

Also have a dongle here with silicon labs cp210 in it that works with the chipkin cas modbus scanner but not modscan32.

Both software packages work with a real serial port and a rs485 adapter. Results above vary per user too, some can't get any to work at all.

Prolific.com.tw has been down for weeks but of course now that I will be busy it is back up so I don't have time to test the updated drivers if there are any.

Have three other dongles here that really aren't any better and mostly have the same chipsets anyway.
 
There are some 10Gb Ethernet NICs on ebay for $120 right now. If I could get four of those and even 4 actual 10Gb switch ports, I'd be thinking long and hard about finding a place for iSCSI instead of local drives.
 
Mercutio said:
There are some 10Gb Ethernet NICs on ebay for $120 right now. If I could get four of those and even 4 actual 10Gb switch ports, I'd be thinking long and hard about finding a place for iSCSI instead of local drives.
Click to expand...

That sounds like a nice deal. I'm not familiar with the CX4 cable, but it looks limited to 15 meters. Would you be running those as a direct connect from machine to machine?
 
Thinking of ddrueding's setup, I think I'd want 4 switched ports to support three servers in some kind of clustered configuration (is that Vsphere?) plus a backend storage server containing the guests. Isn't the point of getting all the VMware stuff in a high availability environment so that the guests can be reallocated dynamically to hardware that's actually operating?

iSCSI or a SAN or something would be a requirement for that kind of setup to provide the storage pool and I'd hope that 10Gb ethernet is fast enough to eliminate a bottleneck for access to that storage.

Unfortunately, it looks like affordable 10Gb switch ports are a flight of fancy all on their own. I guess he could just buy six NICs and run point to point like you were suggesting but... yuck.
 
ddrueding said:
I don't bother with the iSCSI backend, all the disks are local. The simple description of the current problem is just flaky hardware with the simple solution being better and more hardware.

Of course, if I wanted to implement systems and management capabilities that would prevent a flaky machine or two from becoming an all-nighter, VMWare has me covered. I just feel that the additional costs in licenses for the machines and training for me would get us very little at our current size. Perhaps the risk of an all-nighter every 3 years is the cost effective solution.

I'm thinking of just going with 2n redundancy instead of n+1 for the hardware and doubling down on backups via faster disks and a dedicated network.
Click to expand...

You are certainly battling that line of cost of license and increased complexity in the environment to reduce the limited downtime situations. Given the cost of licenses, it's probably cheaper for you to buy one or two more machines as spares since it seems like the business can tolerate some downtime. Centralizing your storage could help make it easier to manage the backups, but it comes with the risk of all your eggs in one basket. At that point the servers really become the expendable commodity and you would begin to rely on their vmotion feature and implement HA and DRS, and possibly deploy VMs with fault tolerance. From a management standpoint, it's been truly wonderful to be able to right click on a server, set it into maintenance mode, and let it automatically push off all the VMs onto other servers and balance the load. I then bring it down, install new memory, power it back on, exit maintenance mode, and then it rebalanced the load. None of my teammates even knew that five machines got upgraded. :-)
 
Mercutio said:
Thinking of ddrueding's setup, I think I'd want 4 switched ports to support three servers in some kind of clustered configuration (is that Vsphere?) plus a backend storage server containing the guests. Isn't the point of getting all the VMware stuff in a high availability environment so that the guests can be reallocated dynamically to hardware that's actually operating?

iSCSI or a SAN or something would be a requirement for that kind of setup to provide the storage pool and I'd hope that 10Gb ethernet is fast enough to eliminate a bottleneck for access to that storage.

Unfortunately, it looks like affordable 10Gb switch ports are a flight of fancy all on their own. I guess he could just buy six NICs and run point to point like you were suggesting but... yuck.
Click to expand...

Yes, the cluster (datacenter) would require running the vCenter software product (you can use a VM, and I believe they give you a 60-day trial before you need to reinstall). You would then take your vSphere (ESX/ESXi) servers and add them into vCenter (which is painless). From within vCenter, you can then go through each ESX server and add your shared datastore, which in your case would be an iSCSI device. As long as it's mapped correctly to all your ESX servers, vCenter will issue a storage rescan and you will see the same datastore name listed on all ESX servers.

Once you have that configured, you will then need to define a private network with a special type known as vmotion. Each of your ESX servers will use a second dedicated network card that will be for vmotion. GigE is acceptable when you're managing only a few dozen VMs (and you can team multiple GigE links if needed). This port group will need to be the same on each ESX machine. You will assign a private IP to each system so that they can communicate. After that, you would then use any other NICs for your public network interface. You can either define them separately on each ESX (making sure they use the same name), or look into a distributed switch configuration. You can use one or more vlan in a port group or just use whatever IP range you manage on your own network.

Once that's configured, vmotioning is seamless between machines. Depending on the amount of memory in use and how busy it is, they can migrate in 30-60 seconds.

Our typical configuration is 4 x GigE NICs. Two are teamed for public, the other two are teamed for private. We have two GigE cisco switches and for the private side, I have a single crossover running between them on the private network for fail-over. I added some pictures to show what the config would look like.
 
Yeah, I don't think I need that level of complexity. At the moment I have the VMs on one ESX server backing up to the other ESX servers, so a once-every-three-years hardware failure simply means firing up the backup that already exists on the other server.

It isn't automated, and there is downtime (couple minutes), but it is so much simpler.
 
1. Put noisy 1U server on your desk
2. Use that as an excuse to put on your noise cancelling headphones and close your door
3. Use that as an excuse to ignore everyone and all phone calls all day

Awesome.
 
The rates in Canada would certainly increase if I would visit Canada. But I have no plan to go to Canada any time soon. It's a nice country, but it's full of Canadians.

BTW, I don't hate the CityK variety (nice guy, nothing against him). It's the John Baird kind, very common in the West, that makes me lose my temper and go berserk.
 
ddrueding said:
1. Put noisy 1U server on your desk
Click to expand...


I'm actually doing something similar. One of our classrooms is completely unusable during the day because of noise from the middle school that's on the other side of wall.

So now I have a dedicated space and explicit approval to work on loud-ass rackmount equipment in that room. I used to have to wait until after business hours before I could get in to that stuff.

I also got a big thumbs up to actually run the 300W sub for the audio system I'm using in my office and/or for my media players class, something I'd previously refrained from doing.

Needless to say, this is a mature and measured response to the shit the school started in the first place.

Good thing no one really expects me to answer the phone anyway.
 
You really do have a crazy work environment. Is this school a public or private school?
 
Handruin said:
You really do have a crazy work environment. Is this school a public or private school?
Click to expand...

IIRC it's a charter school. At this point in time Merc you may want to deploy Santilli's Klipschorns, arranged so the sweet spot is beyond your walls and into the adjoining room in question. With 105db/per watt efficiency, 10W should be sufficient for desired effect.
 
A deal

25% off at ComplyFoam.com

What happened to October?

We don't know about you, but we have been busy with concerts, trade shows, enjoying new music, and updating the list of compatible earphones on our website.

Halloween is almost here, so we are treating you with a 25% discount for all purchases through the eve of Halloween (10/30). Use this code during the checkout process to get the discount: ComplyOct25

...ALSO...

Check out our Facebook giveaway going on now...
Click to expand...
 
Those are different. I see they make them for the iPhone ear buds. Do you have a set of these? What do they offer for you?

Speaking of facebook giveaway deals. I've been signing up trying to win a pair of Grado PS500 headphones. Today is the last day for their current raffle.
 
I have a set on my triplefis. Basically these are foamy ear plugs with a tube in the middle for the IEM. They offer similar noise blocking to regular OSHA 26db foamy ear plugs, maybe more. About as comfortable as foamy ear plugs too.
 
Pradeep said:
IIRC it's a charter school.
Click to expand...

Charter school so... both public and private, I guess. The school does things like block off access to our parking lot for about two hours a day, including times when our students and customers would normally be trying to visit or leave, and use an intercom system throughout the day that is audible through our offices. I'm told that the kids wander up to our windows and stare in while they're at recess or in PE classes (I don't have a window or teach in a room with one), which pretty much everyone finds at least somewhat annoying.

On the other hand, we still have a long term lease at a rate far below market rate for the space and to date neither legal action nor settlement negotiations have led to any remedy for the situation. It's ridiculous but that's where things are.
 
MaxBurn said:
I have a set on my triplefis. Basically these are foamy ear plugs with a tube in the middle for the IEM. They offer similar noise blocking to regular OSHA 26db foamy ear plugs, maybe more. About as comfortable as foamy ear plugs too.
Click to expand...
They look similar to the tips that come with the newer Shure phones. link

The Shure guys gave me some free at CES when they introduced the SE models and I've been using them on my E4c's.
 
That and they seem to come through windows update with the rest of the patches with no fanfare.
 
Stereodude said:
They look similar to the tips that come with the newer Shure phones. link

The Shure guys gave me some free at CES when they introduced the SE models and I've been using them on my E4c's.
Click to expand...

The complyfoam tips actually shipped with my triplefis (6 pairs in the box). They are apparently the new standard. I love them.
 
You must log in or register to reply here.
Back
Top