Marginally OT: Gb SOHO switch recommendations

Same, but they don't vary. My socks are always black (the older shade gets thrown away so they don't get mixed in). Slacks are more expensive, so take a while longer but they're all replaced within a few months.

All of my shirts are button down, long-sleeve. I prefer Oxfords but they're not always available. I wear jeans on the weekend (with long-sleeve shirts) but not to work.

Navy/black gets me in trouble, so no navy!

Yup. I'm a nice guy, so I make it easier on my wife. ;-)

I always catch myself later in the day, then worry for the rest of the day about getting that lecture.

That's why I like Oxfords. They tend to wrinkle less. Some cotton, even advertised as wrinkle-free, is a mess. If I asked SWMBO to iron I'd wear it (the iron, not the shirt).

I have had great success with Zonet and Trendnet. Have several of each under 24/7/365 use and zero problems.

Dave

Jumping from black to white to beige to ... makes me feel a tiny *less* bit "predictable" :> (a VERY tiny bit less!)

Exactly. I think if you have a variety of colors, styles, then you are more likely to end up with an "old" blue sock paired with a *new* blue sock. Ignoring the potential color differences, you then find yourself saying, "Gee, *this* sock looks pretty worn... OTOH, it's *mate* looks pretty

*new*! If I discard the one and keep the other, then I'm stuck with an orphan sock -- until I can find another similar color that 'looks pretty worn'..." Too much thinking for something that should be a no-brainer!

I alternate between a couple pair (jeans). When the knees, crotch, bottoms split/fray on the "heavy use" jeans, I toss them out and take the "clean/newish" pair and demote them to "daily (ab)use" status. (Always have to keep one pair of newish looking jeans to guard against "I won't be seen in public with you looking like that!" :> )

I work out of the house so dress exclusively for comfort. If I want to walk away from "work-work" and do some *yard* work, I don't want to think about "do I need to change into clothes more appropriate for digging?"

Just don't use navy/black for anything that you need to wear *two* of at any given time! E.g., forbidden for socks cuz there is NO WAY to figure out if you've got a matched set on -- until your shoes are laced and someone else is staring at your MISMATCHED socks! :<

OTOH, you only need to wear one *shirt* at a time. If I *thought* I had put on a black shirt only to later discover it was really navy, I wouldn't care -- it's not like I'm trying to match a black shirt to black pants, etc.

I'm a lazy guy -- I make it easy on myself! :> (she does her own, as do I)

Yeah. Amazingly stressful. Akin to being a kid and forgetting to brush your teeth and afraid someone is going to notice something *stuck* in them! :<

When I wore dress shirts, I used to iron them, starch the collars, etc. Ditto for a nice crease in the pants! That's what got me to the point of deciding that there were more important/useful ways for me to use my time! :<

I'm thrilled I wasn't born female! I don't think I could ever deal with the "appearance pressure"!! :-/

I really like the no-iron Brooks Bros shirts. Not too dear when on sale. Socks from Costco or whatever, all black except for a few white sports socks.. something about black socks and shorts is definitely a no-no. ;-) Bass Wilton or Rockport (for walking) dress shoes.

This is why the advertisers go for kids in their early adulthood years, once you settle on stuff that works, there's little reason to change as long as you can buy it at a more-or-less reasonable price.

What changes with the "green" approach? Do ports idle at lower power levels? Unused ports use less? Or, is just the entire fabric made using different technology?

I.e., does a "green" switch supporting all ports running flat out use *less* than a non-green switch in the same configuration? As such, it relies on exploiting usage patterns (admittedly a viable and effective option!) to achieve its savings.

Yup. I would never have noticed the problem (I was working in another room "talking" to a machine in the office) had it not been for the dramatic change in performance!

This was followed by the intermittent "disappearance" of SMB shares (though followup attempts to verify them always yielded positive results!)

Finally, I walked into the office to catch the switch with all lamps illuminated (e.g., like the POST sequence) and watched it go through "aperiodic" cycles like this. Then, wondering how the links worked AT ALL!! :>

No. There are no Telco connections anywhere in the house (with the exception of one specific location).

When I wired the house, I went to extra lengths to ensure it would be difficult to confuse the telco connections with the network ones -- fearing this sort of screwup (potentially worse with PoE).

I believe the biggest change is that banks (usually 4 ports) that don't have anything plugged into them (or at least powered on) have their associated circuitry completely shut off. (Presumably it has to be powered back up occasionally to detect someone plugging in a cable, but at a low duty cycle that's nothing.)

No, I don't believe so... or at least in that case I think the savings are minimal.

Yes, that's my understanding.

---Joel

I suspect there is a separation, in the minds of the marketeers, between "casual" and "serious" users -- based largely on numbers of ports!

I.e., someone looking for a 5 or 8 port switch probably doesn't have the same sort of investment (i.e., more cost conscious) than someone looking for 24 or 48 ports. E.g., my 24 port switch has been running without a hiccup for a long time (OTOH, I could probably buy 10 or 15 eight-port switches for equivalent dollars!)

I checked the wall wart (under load) and it looks fine. UPS isn't acting up (I thought, perhaps, it might have been switching in/out without my noticing). So, next bet is to recap the switch and see if that makes a difference (I've learned it isn't worth the time to drag out a 'scope

*or* LCR meter to check things... just replace the caps and see if the symptoms change! :> Only costs a buck or two and a few minutes with the Leister...)

OK, this looks like just a band-aid approach -- i.e., most existing chips implement 4 at a time so they are just "enable/disable" the entire device? I.e., it doesn;t look like any serious redesign of the silicon (?)

OK.

It would be nice if the architecture's "hacks" were exposed to the end user, then. I.e., silly for me to use ports 1, 5, 9... (leaving the others idle/unused most of the time) thereby almost guaranteeing the switch can never conserve.

OTOH, explaining this to Joe Novice would require some overly simplified scheme: "use ports in the order they are numbered", etc.

That seems to be the case with much kit. I've never bothered trying to salvage switches, previously, as I figured they were practically disposable. And, coupled with their high duty cycle (i.e., ~9000 PoHr/year) it seemed inevitable that even "quality" caps would "go toasty".

I'll open it up later this evening and recap it (there can't be *that* many caps in the thing!) and see where that leaves me...

The datasheet for mine says that they reduce power to unused ports, reduce power to ports connected to devices which are off, and reduce power based on cable length. "Up to 70% power reduction", whatever that means.

Sure it isn't just the wall wart dying?

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

"This week only! Save up to 70% on select models in our dealer inventory! No reasonable offer will be refused!"

Yours appears to be easier for a user to relate to -- without even understanding the issues involved. "Unused capacity results in a direct savings"

Naw, I checked that. I even checked the UPS into which it was plugged on the off chance that the UPS was "switching" without tell anyone.

Chances are, it will be bad caps or something got "a bit too hot" and went south.

The fact that it works perfectly *except* for these "fits" is the only incentive to open it up (previous switch failures were more "thorough" -- "completely dead").

I vary the shirt color. ;-)

;-)

I do have more than one set of socks; Winter/warm/Summer/work. White gets thrown in for work or warm.

I have three pair of jeans (weekend + wash) and seven or eight pairs of slacks for work, all the same black and same brand/style. It also takes a while to replace them because stores only carry one or two of each size.

I work for a large company with a fairly rigid dress code (and worked for IBM for >30 years). The last job I had was incredibly lax but I still wore the same clothes.

But navy pants and a green shirt looks like mommy should still be dressing you.

Division of labor. It's more efficient.

Amen to that! I know women who never wear less than $500 in clothes outside of the house. Then there's the hair and makeup thing! Nope, I'm with you.

I do it. Tough if it offends the "no white after Labor Day" crowd. I have white athletic socks, too, but I can't be bothered thinking about which to wear.

I have really odd sized feet, so about the only shoes that fit are New Balance. They have one style in black that is passable for work and I have a white pair for weekends.

...until the bastards kill the style that fits. "New and Improved" never is.

Ah, well... I consider *shorts* a "no-no" (exposed skin being something we frown upon in skin cancer capital of the country :< ) Much easier to cover the legs with pants than to mess around with all that gunk you'd have to rub over them (think about the hair on your legs being *stuck* to them with an "adhesive"... and how that feels as you flex them!)

My current walking shoes are NB. But, an old style no longer produced. With 837,426.2091 different styles to choose from, I suspect I will never find The Right One (assuming they still

*make* something that I am comfortable with!)

Exactly! That's the boat I'm in with my walking shoes! :< You can put up with a shirt that's an odd color. Or, a bit too large. You can put up with pants that are a bit too long. Or short.

But shoes *have* to fit and be comfortable. Especially if you are in them for any amount of time or exertion (I take a brisk

4mi walk, daily... shoes that don't fit well quickly eat up my feet!)

Why can't folks stop dicking with things? *How* is it improved FROM *MY* STANDPOINT? Yes, I'm sure it is improved because you have found a way to produce it more cheaply. But how does that translate into an improvement for *me*? And, you want $50/pair -- or, $50/shoe! -- for this??

I pretty much stick to black, navy and white. Perspiration is more apparent on other colors. :<

Ah, for me, they are all identical. Buy a few dozen at a time. If just a few wear out, you can cull them. But, there comes a point where they *all* start wearing out. Time to buy a new batch and toss all of the remaining old ones. (I tend to abuse most of my clothing... i.e., walking down the street in "stocking feet" -- then being met at the door by The Frowny Face. I just don't seem to have much "common sense" when it comes to things like that! :> )

In my case, it's usually being an "odd" (literally) size that makes things tough. I'm now in an "even" size which makes the selection easier! (no idea why "odd" is so much harder to find...)

My boss at my last "regular" job used to remind me *daily* to wear a tie. Years later, I *still* never remembered! :>

[OTOH, if I was representing the company "somewhere", I'd wear a three piece suit, etc. But, every day clothes should never be fancier than casual slacks and a "nice shirt"]

Ditto regarding the green. There's just something *wrong* with green! Whether its shirts, pants, socks, cars, houses, etc.!

But, with just three colors for shirts and "always jeans", its hard to end up with a "bad combination"! :> *Boring*, yes. But never mismatched!

Perhaps. We split some things along rigid lines. E.g., I do the baking -- but never clean the counters afterwards! I figure that;s the "cost" she has to bear for the baked goods! (I don't eat sweets -- just make them!)

Or 85 pair of shoes! I now have *three* pair of shoes and it leaves me always looking for the *right* pair (walking, dress, yardwork). I have no idea how I would ever keep track of the appropriateness of any greater number than that!

Years ago, I would wear boots, exclusively. Made it really easy to decide which pair of shoes I was going to wear (the everyday boots or the ones that haven't yet been delegated to everyday use -- "dress boots") and completely eliminated the sock issue (navy+black, pink+yellow, bare feet, etc. -- no one was the wiser! :> )

Unfortunately, boots suck when it's perpetually warm outside

*and* when you do a lot of walking.

Well, after checking the UPS and wall wart (see posts elsewhere, this thread), I made some time to open this puppy up, tonight.

Rather than waste time *measuring* and *testing*, I just unilaterally picked the caps most likely to be causing problems (no visible signs of bulging or leaking electrolyte) and replaced them -- *just* them!

Switch has been running for about an hour now without me being able to "catch" it in the act of resetting itself (which is not to say that it

*isn't* doing so!).

I'll power up another host, tomorrow (something with reliable and PREDICTABLE network characteristics) and just 'ping -i0.5 -q ' -- leaving it run for hours!

Then, poke it with a kill -SIGINFO from time to time to check the accumulated statistics -- watching the number of dropped packets (should be NONE) and maximum RTT (on the assumption that a switch reset will cause packets to be dropped *while* the switch is recovering and, with a reasonably fast ping interval, that number should be significant over the course of hours -- even MINUTES!)

If these results suggest the switch is working, again, I'll open it up (again) and replace the remaining caps (since they are probably also degraded but not yet symptomatic)

I also need to make a note of the date at which the repair was made so I can see how long *these* caps last!

All in all, a productive evening:

- switch appears to be working

- repaired an LCD TV (fuses, caps and FETs on inverters)

- prepped a VCR for troubleshooting later this week

Tomorrow, repair the driver's side checker on her car...

Ahhhh, life is good! ;-)

I think it uses later, lower power chipset.

Well I have a mix of 100Mbps and only three 1Gbps machines hooked to the thing. Most of the bulk traffic is from one to another box during data rate testing last year for many hours at a time[1], in the background is slower Internet access and so on for most boxes.

[1] Loading up and testing six 1TB drives in a RAID6 box for 4TB online store. Built before drive prices went up last year.

The Netgear 1Gbps switch ran too hot, before that I had a hot running 100Mbps switch, sold them on ebay. Netgear had a bad name for over-running their power supplies.

Yes, look down at the lights and WTF?

SMB is a bit slow to recover, I'm running mix of Linux and Windows, both SMB and NFS sharing.

There's a story of plugging in "Station Sync", video frame rate signal into BNC network connectors (thin ethernet) that didn't stop the networking, just slowed it some.

Grant.

OK. Different reports (here) seem to indicate different approaches...

But, how did this compare to "green" technology (performance and power dissipation in the same loading)?

In my case, it meant walking into another room. As I first walked in, I caught the tail end of a "reset" sequence and wondered if my eyes had just "played tricks on me" (since the switch's indicators immediately returned to "normal", transfers resumed, etc.).

But, another "reset sequence" shortly followed and I was *sure* of what I had seen. :<

Understood. In my case, I would start a transfer (e.g., copy a "folder" with several GB in it from one machine to another) and walk away. Come back a few minutes later and see a message claiming that the share had disappeared!

*Look* at it, again, and "Gee, it's STILL THERE! WTF??"

Resume the transfer -- watching for a few moments -- then return some time later and it may have finished normally

*or* coughed up the same error message.

Of course, the first thing you think of is a problem with a

*machine* -- did the source of the transfer crash/hiccup? It takes a while to consider the fabric as a possibility!

Ha!

I sorely miss 10Base2 (though not the *speed* limitations!). It was *so* much easier for me to cable than 10[*]BaseT...! (i.e., just string the cable from one host to the next... to the *next* ... to the ...)