Results 1 to 18 of 18

Thread: Linksys SRW224 Switches Randomly Crashing

  1. #1
    Member c4p0ne's Avatar
    Join Date
    Sep 2002
    Location
    Classified
    Posts
    95

    Arrow Linksys SRW224 Switches Randomly Crashing

    Honestly I'm at a loss for words. I just can't seem to figure out what's the root of this problem. The network is setup as follows:

    40 Machines all running Windows XP Pro (38 Clients, 2 Servers).
    2 Linksys SRW224 Switches w/ Web Management.
    The 38 Client Machines are using VIA Rhine II Network Adapters (built-in)
    The 2 Machines acting as servers are using different network adapters.
    The Linksys Switches are connected to eachother via the Gigabit Ports.

    The environment is an Internet Cafe In Greece. What's happening (seemingly RANDOM) is that out of nowhere, the whole network just DIES. I mean literally everything goes DEAD. The switches seem to freeze up and nothing is pingable or accessible either internal or external. IT "SEEMS" to be more common when ALL of the client machines are in use. It very RARELY happens when the shop has normal volume of customers.

    For instance over the weekend, every single PC was taken and everyone was playing games (9 out of 10 PC's were actively playing Lineage II). Suddenly the internet cafe management software begins to act erratically (the first indications that I will be feeling the need to kill myself in a minute or so)... Certain machines start losing connections and then re-gaining them again. Sure enough after 2 minutes everything goes totally BLACK. Everyone gets disconnected and the entire shop empties with angry gamers withing minutes afterwards.

    A loss of 80 Euro's per hour of downtime. That is NOT good. At this point I'm not even ruling out sabotage such as a possible ARP Poisoning attack etc.. However I still feel that it is SOMETHING ELSE. Can anyone here shed some light on this freakish situation? After rebooting the switches as well as ALL 40 machines everything slowly goes back to normal and connections start getting established again. Broadast storm? ARP storm/poisoning? Electrical problem? Wiring problem? Switch misconfiguration/overload? WTF i'm becoming bald trying to figure this crap out.

    There are massive efforts on the part of the internet's corporate owners to try to direct it to become a technique of marginalisation and control. --Noam Chomsky
    ---

  2. #2
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    I've been installing quite a few of those Linksys/Cisco switches lately, and have an excellent experience with them so far.

    Today, I have 2x of them in my Jeep, to go install on a network to replace some DLink DES 32 port switches, one of which started crashing a lot recently. Funny thing is, this existing network has quite a few clone built computers with that (Ugh) Via Rhine 2 adapter....I'm pushing this client to replace those machines as quickly as possible. Soyo motherboards I think. In the past, I've run across issues with these Via NICs.

    I don't remember what firmware is the latest...I'll post when I'm onsite, but is yours up to date?

    The giga-cable uplinking the two, is it CAT 5e or CAT 6 or CAT 7? Home made or factory made?

    Any cables home made? Or all factory made?

    Switches on battery backup?

    Adequate ventilation? Under full load, they're not stuffed in some cramped area overheating?

    You changed the default LAN IP of one of the switches, correct? So they're not both the same IP.

    Disable MAC ageing?
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  3. #3
    Member c4p0ne's Avatar
    Join Date
    Sep 2002
    Location
    Classified
    Posts
    95

    Unhappy

    Quote Originally Posted by YeOldeStonecat
    I've been installing quite a few of those Linksys/Cisco switches lately, and have an excellent experience with them so far.

    Today, I have 2x of them in my Jeep, to go install on a network to replace some DLink DES 32 port switches, one of which started crashing a lot recently. Funny thing is, this existing network has quite a few clone built computers with that (Ugh) Via Rhine 2 adapter....I'm pushing this client to replace those machines as quickly as possible. Soyo motherboards I think. In the past, I've run across issues with these Via NICs.

    I don't remember what firmware is the latest...I'll post when I'm onsite, but is yours up to date?

    The giga-cable uplinking the two, is it CAT 5e or CAT 6 or CAT 7? Home made or factory made?

    Any cables home made? Or all factory made?

    Switches on battery backup?

    Adequate ventilation? Under full load, they're not stuffed in some cramped area overheating?

    You changed the default LAN IP of one of the switches, correct? So they're not both the same IP.

    Disable MAC ageing?
    1. The switches are linked by a CAT5e cable on the 1st gigabit port on the 1st switch to the 1st gigabit port on the 2nd switch. Factory made of course.

    2. The entire case that the switches are in IS INDEED CRAMPED, however the thing's got 2 industrial (i would say) strength fans working to suck decent amounts of heat from outta that little puppy. Everything with the exception of the 2 fans inside the case is on a UPS. The switches have about 2 inches of space between them.

    3. Unfortunately ALL the cables running from the 38 Client machines + the 2 machines acting as servers are gayly home-made. You know, this is a classic scenario where the business men are trying to "cut corners" saving money the WRONG WAY by hiring (apparently) what seem to be kindergarteners to build their production network; then forcing another admin (me) to deal with the problems of the previous (now vanished into thin air) idiots.

    4. Yes both switches have different IP's @ 192.168.1.254 & 192.168.1.253.


    5. I'm not sure what state I left the MAC ageing settings on, however I do believe the last I checked they were enabled @ 150.

    So you think that the VIA Rhine II sucks eh? Wanna know a secret? I'm starting to think so too. I mean the MAC-prefixes of this vendor don't even appear to be in any of the well-known reference databases. Crap. Wondering if it is a good idea to force ALL 40 machines to 10/100F instead of auto-negotiate and then doing the same for the switches...

    It COULD be that someone is coming in (rivals are vicious here) and firing up some kind of ARP attack by poisoning traffic between the gateway and all the machines thereby causing a tremendous amount of stress bringing the whole thing to it's knees, however this is unlikely as most of their clients are children playing LAN games.

    If it is the wiring issue I wonder what I should charge this guy because right now, testing/rewiring the entire infrastructure is just simply something I can't do since I have at least 3 other netcafe's in Athens & Pireus to fly back and fourth to on a weekly basis.
    There are massive efforts on the part of the internet's corporate owners to try to direct it to become a technique of marginalisation and control. --Noam Chomsky
    ---

  4. #4
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    1) OK, that's fine. Today I did mine using Giga-2 on each for the uplink, Belkin Cat5e 1 foot cable to uplink them. I turned on QoS to high for that port, the internet port to the Cisco PIX 501 (port 24 on second switch), and both of the first giga ports.

    2) So you're thinking adequate airflow completely across the switches? Cuz one of the first things that came to mind..you mentioned "when under full load". Makes me think "heat". Anything else you can do to possibly open up this area more, "just to see if it helps"? Don't forget, those air holes are in the sides. Rubber pads put on the bottoms, so they don't sit flat on top of whatever they're on? Adequate space in between them? And the location? They're not located like...right under some flourescent light fixture, or something else that would emit EMI, right?

    3) Ugh...home made cables. Unless you know they were made by a professional who tested them with a HIGH QUALITY tester, and not just some cheapo link light test....then I'd consider this an issue. Todays onsite, a Hospice I recently picked up a few weeks ago as a client, the joker who did the home made wirings originally...UGH what a poor job. Reterminating them is on my short list of things to do. Home made wires made by people who "think" they can do CAT wiring has caused me so many headaches in my career. The percentage of people who can actually hand make them and have them run problem free is very small. Now let me further state..that when working with giga-over-copper...tolerances are even tighter..home made gigabit wiring...ACK.

    4) OK good. Oh, the firmware on the shipping ones I received this week is version 1.3. None available for download at the Linksys site, however, a few weeks ago a Cisco tech e-mailed me a newer firmware when I was troubleshooting some odd issues at another network I was upgrading, which ended up not being the switch anyways. I'll see what version that is, can e-mail it to ya if ya like.

    5) Might try disabling MAC ageing just for testing..who knows with those Via NICs. Yes I abhor Via products, NICs, chipsets, etc.

    Now for some of my experience today, this medical nursing scheduling database that the Hospice runs...is on a Dell PowerEdge Xeon server, with a gigabit NIC. So is the DC, and the several other servers are just on 10/100 NICs. My initial design was use a gigaport on each switch to uplink each other, and have the two Dell servers in the other two gigaports. Well this scheduling software, Suncoast, runs on an oddball database engine called Progressive. Looks similar to VFoxPro to me..which is normally a pig on networks, very heavy traffic. Well, this software freaked out when running on gigabit. Every couple of mines, all clients would get get bunked out of it with "write" errors. Took that server, put it into a regular 10/100 port on the switch..all was well. I was scratching my head for a while...either that NIC was going..and couldn't handle gigabit, or the software itself just didn't like it (which doesn't make sense to me, I've never encountered that before...software that didn't like running at a faster speed)

    ///shrugs shoulders///

    Anyways, you mention you had 2 servers plugged into the gigaports...just for yucks, can you plug them into 10/100 ports? Maybe they can't handle the giga speeds.
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  5. #5
    Member c4p0ne's Avatar
    Join Date
    Sep 2002
    Location
    Classified
    Posts
    95

    Arrow

    No, none of the machines are plugged into the giga ports, only the 2 switches are connected via gigabit to eachother.. All the machines in the place have 10/100 cards (or built in) in them. Another thing that just popped into my head is that they're running an Enterprise Edition of DeepFreeze, which sends out "heartbeats" to the central console every 30 seconds according to the white-paper on the software. This is in the form of some kind of broadcast. All though I DONT think that this is the actual problem (since another cafe I work with is running the same with OVER 100 machines with NO issues) it MIGHT be contributing to it.

    What I'll do is disable the networking capability of DF to see if this helps... Perhaps I'll install a "pro" version with no networking and just mass-control it through IPC$. Oh, and you said that the firmware you saw for the Linksys SRW's is at v1.3?? Hmm, these are still at v1.1 last time I checked but I can't seem to find a later version on linksys site....
    There are massive efforts on the part of the internet's corporate owners to try to direct it to become a technique of marginalisation and control. --Noam Chomsky
    ---

  6. #6
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    Heartbeats...nah I doubt it's that, that's actually very little traffic. Which I-Cafe software are you running? Nothing that's actually proxy, correct? Although I wouldn't see that being any factor at all.

    PM me your e-mail addy if you wish, I'll mail ya the .bin file.

    What kind of internet connection? And how connected?
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  7. #7
    Member c4p0ne's Avatar
    Join Date
    Sep 2002
    Location
    Classified
    Posts
    95
    Well it's DSL over ISDN crap; PPPoA. Greece is really behind in terms of this stuff but anyway, I wish the switches were SNMP enabled because I can guarantee that this dumb issue would be solved by now but I guess you win some, you lose some. I'm gonna back up there on Sunday and try to research this issue some more. This is an Arab guy that owns this thing and lemme tell you he's so tight with money, that once he had his birthday 3 times in one month to postpone paying me. Eventually I got every cent but as far as this "problem" is concerned, it's PERSONAL.

    I really can't rest until I find out what in the HELL is causing this, even if it takes months. Say BTW how did you get your hands on that firmware again?? the v1.3 that you mentioned? my email: pr0active-at-hotmail-dot-com you mind firing that thing over my way? BTW These are EUROPEAN versions (not sure if theres a difference other then the 220V power supply) of the Linksys SRW224 switches.
    There are massive efforts on the part of the internet's corporate owners to try to direct it to become a technique of marginalisation and control. --Noam Chomsky
    ---

  8. #8
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    Quote Originally Posted by c4p0ne
    Say BTW how did you get your hands on that firmware again?? the v1.3 that you mentioned?
    I have a reseller account with Linksys/Cisco..."Tech to tech support" package I signed up for years ago. Comes in handy, pretty much a direct call...1x push button menu to enter account number, then a quick "live body pickup".
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  9. #9
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    1.3 just went on its way to ya, if I can get anything newer I'll also forward it.
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  10. #10

    copy

    I have this IDENTICAL problem! Often it isn't all of the devices loosing connectivity until they're reset, but half. It's especially annoying that the APC unit keeps doing this, as to reboot that takes out servers in the office it's connected to.

    Could I, pretty please, have a copy of the firmware to try?

    With regards to the setup, we have one of the switches with half a dozen servers, workstations, ip phones, network cameras, printers etc. It's happening when at LOW LOAD for us here. All factory ethernet cables.

  11. #11
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    Quote Originally Posted by MrBobbyEwing
    It's especially annoying that the APC unit keeps doing this, as to reboot that takes out servers in the office it's connected to.
    I don't get what you mean by this. The APC unit keeps dropping power to the switches?

    PM me your e-mail addy, I'll send you the firmware...I think I still have it.
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  12. #12
    Junior Member
    Join Date
    Jul 2005
    Posts
    2
    I also have a client whose SRW224 has been freezing up. Every once in a while all network traffic just dies. The lights on the switch look normal, but you can't ping anything, etc. Rebooting the switch takes care of the problem.

    The SRW224 has been RMA'd, but it didn't fix the problem. After moving them to an HP ProCurve switch they had before, the problem went away. Unfortunately the HP switch is too small.

    It is an office environment with 20 or so PCs. All the cables are factory-made CAT5e. The switch is in a cabinet with the server and everything else, but there is plenty of space. I suppose it could be overheating, but the load is not heavy, and the ProCurve switch and all other devices are fine. I'm not sure what version the firmware is.

    I'd hate to be a bother, Stonecat, but perhaps I could snag the firmware from you as well? I intend to continue researching the problem, but I have a feeling the SRW224 may simply be a sub-par product. I prefer ProCurves myself (especially for the management interface, once you update the firmware,) but they are more expensive.
    Last edited by xannonite; 07-07-05 at 01:21 PM.

  13. #13
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    Quote Originally Posted by xannonite
    I'd hate to be a bother, Stonecat, but perhaps I could snag the firmware from you as well?
    Not a problem, PM me your e-mail address and I'll beam it out to ya. (just give me a day or so, I'm about to leave the office and head out on the road for the rest of the afternoon for an onsite) Actually this thread is a bit old, I'm sure they have a newer version out, I'll get in contact with the Cisco VAR tech and try to get something newer.

    I have a few of these out in the field now, I love them.
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  14. #14
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    Firmware sent today Xannonite.
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

  15. #15
    Junior Member
    Join Date
    Jul 2005
    Posts
    2
    Quote Originally Posted by YeOldeStonecat
    Firmware sent today Xannonite.
    Thanks. Unfortunately when I went in to update it I saw it was already v1.3. I did disable MAC table aging while I was in there. Ah well.

  16. #16

    Cool I to have felt your pain...

    Quote Originally Posted by c4p0ne View Post
    Honestly I'm at a loss for words. I just can't seem to figure out what's the root of this problem. The network is setup as follows:

    40 Machines all running Windows XP Pro (38 Clients, 2 Servers).
    2 Linksys SRW224 Switches w/ Web Management.
    The 38 Client Machines are using VIA Rhine II Network Adapters (built-in)
    The 2 Machines acting as servers are using different network adapters.
    The Linksys Switches are connected to eachother via the Gigabit Ports.

    The environment is an Internet Cafe In Greece. What's happening (seemingly RANDOM) is that out of nowhere, the whole network just DIES. I mean literally everything goes DEAD. The switches seem to freeze up and nothing is pingable or accessible either internal or external. IT "SEEMS" to be more common when ALL of the client machines are in use. It very RARELY happens when the shop has normal volume of customers.

    For instance over the weekend, every single PC was taken and everyone was playing games (9 out of 10 PC's were actively playing Lineage II). Suddenly the internet cafe management software begins to act erratically (the first indications that I will be feeling the need to kill myself in a minute or so)... Certain machines start losing connections and then re-gaining them again. Sure enough after 2 minutes everything goes totally BLACK. Everyone gets disconnected and the entire shop empties with angry gamers withing minutes afterwards.

    A loss of 80 Euro's per hour of downtime. That is NOT good. At this point I'm not even ruling out sabotage such as a possible ARP Poisoning attack etc.. However I still feel that it is SOMETHING ELSE. Can anyone here shed some light on this freakish situation? After rebooting the switches as well as ALL 40 machines everything slowly goes back to normal and connections start getting established again. Broadast storm? ARP storm/poisoning? Electrical problem? Wiring problem? Switch misconfiguration/overload? WTF i'm becoming bald trying to figure this crap out.

    I had the same issue with two Linksys switches. They would start to roll off connections then the whole switch would stop responding. I ended up replacing both of them, for this very reason. I know they will do this if there is a machine with excessive CRC Alignment errors on any given port. Check it out, hopefully this helps you out.

  17. #17
    Member c4p0ne's Avatar
    Join Date
    Sep 2002
    Location
    Classified
    Posts
    95

    Thumbs down

    My my my my myyyyy! what a surprise almost a whole 2 years later I see a reply to a really old post! Welp, there's a few things that have changed since then, like they never really figured out what that problem was.. (2 f'ing years) I barely stop by there just to say hi every few months btw I've been all around Athens doing more meaningful work then busting my hump for 2 cheap-a$$ tech-business illiterate Lebanese clowns (I had to sugarcoat it a little so as to sound a little nicer then what I really had in mind).

    So anyway they went ahead and dumped the switches for Linksys (again) Gigabit Web-managed models (the SRW2024's I believe). This did not solve the problem at that time and they were furious. Eventually they found out that it was partly an electrical problem with the F'ing LIGHT DIMMER SYSTEM of ALL things can you believe it? Everytime the lights wen't on upstairs there would be a complete global loss-of-connectivity. But the thing was, ONLY if the lights were "dimmed" to a to a particular setting. All other light settings caused severe slowdowns (but not complete network blackouts).

    So they removed the dimming system and everything suddenly went back to a semi-normal state where things would function within THEIR parameters for what "well" should be (not mine). However this didn't last long. Although they never again had a "complete loss", they started experiencing the severe slowdowns again. Randomly I might add. One could never tell when it was going to happen. When the shop is full? Empty? 5 customers? It didn't matter it would just "happen".

    By that time of course I had already left for reasons of course I just didn't feel compelled to be around a bunch of cheap arab buffoons who cared nothing about feeding their goose but instead slaughtering it on a daily basis looking for non-existent golden eggs. It turns out, you can't help people that don't want to help themselves is true after all. As far as I know, they have these issues to this very day. I'm surprised they still operate actually) at only 2 to 300€/day. When they first opened they were bringing in on average 1200€/daily (never mind the weekends!).

    As you can see the disillusioning of their customers meant nothing to them and their subsequent and still unfathomable stupidity keeps them in a perpetual mental state of retardation asking why they went from making a fortune to peanuts. They STILL don't know and refuse to believe that changing the crap infrastructure of their network and bringing in REAL electricians (in other words spending "more" money) WOULD (not anymore) have made the difference. The problem was solved, it was right there! But nope. "we've already put too much money in this business". One of the morons went and bought a yacht instead of re-investing in a shop that was obviously declining profit wise due to these issues.

    If you ever see jokers like this, take the money and run. I have never been glad to see a business that I worked for go under. Hope they close soon (looks like it). And also, to this day I will never understand how 2 men can run their business KNOWINGLY into the ground and think its going to get better by buying "leather chairs" for the customers. WTF good is a leather chair if you can't check your f'ing email? My God the humanity!
    Last edited by c4p0ne; 11-21-06 at 03:18 PM.
    There are massive efforts on the part of the internet's corporate owners to try to direct it to become a technique of marginalisation and control. --Noam Chomsky
    ---

  18. #18
    Moderator YeOldeStonecat's Avatar
    Join Date
    Jan 2001
    Location
    Somewhere along the shoreline in New England
    Posts
    50,427
    LOL lighting systems eh? Cables must have been run too close the..err..whatever dimming systems use for the control, some rheostat like device. Not unlike the ballasts for flourescent lighting...keep network cables away from them!
    MORNING WOOD Lumber Company
    Guinness for Strength!!!

Similar Threads

  1. $20 Gift Certificate for $100+ purchase of Linksys products
    By CRUM6178 in forum Marketplace - Buy/Sell/Trade/Hot Deals
    Replies: 0
    Last Post: 08-12-04, 10:18 PM
  2. AT&T Broadband and Linksys Cable Router
    By TeamHog in forum General Broadband Forum
    Replies: 1
    Last Post: 12-09-01, 06:51 PM
  3. Beld And Linksys And Toshiba Cable Modem
    By CampRyenWater in forum Networking Forum
    Replies: 0
    Last Post: 07-16-01, 04:56 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •