Re: [c-nsp] WS-C2970G-24TS as access switches
Can anyone tell me if WS-C2970G-24TS practically bad and should be avoided at any cost? If your use case includes any amount of shared access to iscsi, smb/cifs, or nfs and could result in *any* amount of oversubscription between hosts while then I'd suggest the 4948. Dual ac/dc ps, replaceable fan module, $1100 to $1300 ebay style, but 48 copper gige and 4 sfp, and active queue management (with deep enough buffers to matter) ala dynamic buffer limiting. 3x more than I see the 2970 going for, but about 10x the utility. -Tk ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: [c-nsp] WS-C2970G-24TS as access switches
On 12/28/11 11:02 AM, Mike wrote: I was using these for exactly the same reasons stated above. This year, I have had three seperate instances where the switch had to lose power (move, re-work pwr arrangements, etc), and all three times the PSU apparently gave up the ghost and refused to power back up. Nothing 'happened' funny power wise, not zapped or otherwise mistreated in any way. I think these units were of a vintage vulnerable to the bad chineese capacitor problem and I think whatever cap in the psu just went fizzle while it was operating, which would let the units continue running but once it lost power, would prevent a successful full power on start up. This is a very common failure mode with some types of switching power supplies. It is typically a resistor and not a capacitor. We saw a lot of it with the power bricks supplied with Fujitsu DSL modems a few years ago. It's real fun when there's a widespread power outage and customers all over town are down once power is restored. There's a high value resistor, typically in the hundreds of kilo-ohms used to kick-start the switcher. Once it's going, the resistor isn't needed until power is removed and restored. These typically fail open. If the gear is worth salvaging or if it's crucial to get it back online while waiting for a replacement, I typically replace these with a resistor of substantially higher power rating than the original. I was able to find and deploy the rps-675 (redundant power) after being burned this way three times, and it came in damm handy because there was a 4th event (another burned up 2970 psu) and this time the 675 kept it running till I was able to have an orderly replacement and maintinence window (with a 3560). I would reccomend deploying the rps units if you are going to use any cisco products with single power supply, but especially if you're going to be using the 2970's which have proven (in my shop) to be a (literally) dying breed. These power supplies are commodity items from Chinese manufacturers that are used in a variety of gear, not just Cisco switches. You can often Google the part number on the power supply brick itself and find replacements. -- Jay Hennigan - CCIE #7880 - Network Engineering - j...@impulse.net Impulse Internet Service - http://www.impulse.net/ Your local telephone and internet company - 805 884-6323 - WB6RDV ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: [c-nsp] WS-C2970G-24TS as access switches
On 12/29/2011 12:28 AM, Jay Hennigan wrote: On 12/28/11 11:02 AM, Mike wrote: I was using these for exactly the same reasons stated above. This year, I have had three seperate instances where the switch had to lose power (move, re-work pwr arrangements, etc), and all three times the PSU apparently gave up the ghost and refused to power back up. Nothing 'happened' funny power wise, not zapped or otherwise mistreated in any way. I think these units were of a vintage vulnerable to the bad chineese capacitor problem and I think whatever cap in the psu just went fizzle while it was operating, which would let the units continue running but once it lost power, would prevent a successful full power on start up. This is a very common failure mode with some types of switching power supplies. It is typically a resistor and not a capacitor. We saw a lot of it with the power bricks supplied with Fujitsu DSL modems a few years ago. It's real fun when there's a widespread power outage and customers all over town are down once power is restored. In this case with the 2970's, it's definately a capacitor. I don't have my photos now but it was the same cap all 3 times that bulged out and made it obvious it was the problem. These power supplies are commodity items from Chinese manufacturers that are used in a variety of gear, not just Cisco switches. You can often Google the part number on the power supply brick itself and find replacements. I had that idea too but for the 2970 all I was able to find were branded replacements costing 3x the cost of the used rps675... Mike- ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
[c-nsp] WS-C2970G-24TS as access switches
Hi list, Right now looking for new access switches, with 24 copper 1gig ports + 4 SFP ports. We don't really need any fancy features, and from datasheet I can tell WS-C2970G-24TS, fits us very well. I know this is EOL already and replaced with WS-C2960G-24TC-L, which recently replaced with WS-C2960S-24TS-L, but I don't see reason spend more money for features we won't use. Can anyone tell me if WS-C2970G-24TS practically bad and should be avoided at any cost? Thanks ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: [c-nsp] WS-C2970G-24TS as access switches
I've noticed WS-C2970G-24TS actually 1,5U, while other model just 1U. So I wonder if there any overheating issues? On 28/12/11 18:01, Joe Maimon wrote: Works for me in a number of locations. I have seen about one out of 20 go bad. So get some spares. Nikolay Shopik wrote: Hi list, Right now looking for new access switches, with 24 copper 1gig ports + 4 SFP ports. We don't really need any fancy features, and from datasheet I can tell WS-C2970G-24TS, fits us very well. I know this is EOL already and replaced with WS-C2960G-24TC-L, which recently replaced with WS-C2960S-24TS-L, but I don't see reason spend more money for features we won't use. Can anyone tell me if WS-C2970G-24TS practically bad and should be avoided at any cost? Thanks ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/ ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: [c-nsp] WS-C2970G-24TS as access switches
Nikolay Shopik wrote: I've noticed WS-C2970G-24TS actually 1,5U, while other model just 1U. So I wonder if there any overheating issues? The physical design appears to be the same as 3550-12G and some of the 3750. The extra height is taken up by the PS and 2 fans which are on their own shelf, as there is no room behind the mainboard, since the mainboard is deeper. This is 1 more fan then many of the 1U switches have. Overheating hasnt been a problem. They all run from warm to hot to the touch, from the 3500/2950 and up. Your servers will complain well before the switch does. Joe ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: [c-nsp] WS-C2970G-24TS as access switches
The 2970 series also had issues with electrical imbalances. These really need to be chassis grounded where other switches aren't as sensitive. sarcasmNot that anyone ever fails to attach grounding cables to switches./sarcasm If you are running more than one Ethernet cable to the same device from the 2970 you can have issues with the poor handling of imbalances. LR Mack McBride Network Architect -Original Message- From: cisco-nsp-boun...@puck.nether.net [mailto:cisco-nsp-boun...@puck.nether.net] On Behalf Of Joe Maimon Sent: Wednesday, December 28, 2011 8:19 AM To: Nikolay Shopik Cc: cisco-nsp Subject: Re: [c-nsp] WS-C2970G-24TS as access switches Nikolay Shopik wrote: I've noticed WS-C2970G-24TS actually 1,5U, while other model just 1U. So I wonder if there any overheating issues? The physical design appears to be the same as 3550-12G and some of the 3750. The extra height is taken up by the PS and 2 fans which are on their own shelf, as there is no room behind the mainboard, since the mainboard is deeper. This is 1 more fan then many of the 1U switches have. Overheating hasnt been a problem. They all run from warm to hot to the touch, from the 3500/2950 and up. Your servers will complain well before the switch does. Joe ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/ ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: [c-nsp] WS-C2970G-24TS as access switches
On 12/28/2011 05:16 AM, Nikolay Shopik wrote: Hi list, Right now looking for new access switches, with 24 copper 1gig ports + 4 SFP ports. We don't really need any fancy features, and from datasheet I can tell WS-C2970G-24TS, fits us very well. I know this is EOL already and replaced with WS-C2960G-24TC-L, which recently replaced with WS-C2960S-24TS-L, but I don't see reason spend more money for features we won't use. Can anyone tell me if WS-C2970G-24TS practically bad and should be avoided at any cost? I was using these for exactly the same reasons stated above. This year, I have had three seperate instances where the switch had to lose power (move, re-work pwr arrangements, etc), and all three times the PSU apparently gave up the ghost and refused to power back up. Nothing 'happened' funny power wise, not zapped or otherwise mistreated in any way. I think these units were of a vintage vulnerable to the bad chineese capacitor problem and I think whatever cap in the psu just went fizzle while it was operating, which would let the units continue running but once it lost power, would prevent a successful full power on start up. I was able to find and deploy the rps-675 (redundant power) after being burned this way three times, and it came in damm handy because there was a 4th event (another burned up 2970 psu) and this time the 675 kept it running till I was able to have an orderly replacement and maintinence window (with a 3560). I would reccomend deploying the rps units if you are going to use any cisco products with single power supply, but especially if you're going to be using the 2970's which have proven (in my shop) to be a (literally) dying breed. Mike- ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
Re: [c-nsp] WS-C2970G-24TS as access switches
On 28/12/11 23:02, Mike wrote: On 12/28/2011 05:16 AM, Nikolay Shopik wrote: Hi list, Right now looking for new access switches, with 24 copper 1gig ports + 4 SFP ports. We don't really need any fancy features, and from datasheet I can tell WS-C2970G-24TS, fits us very well. I know this is EOL already and replaced with WS-C2960G-24TC-L, which recently replaced with WS-C2960S-24TS-L, but I don't see reason spend more money for features we won't use. Can anyone tell me if WS-C2970G-24TS practically bad and should be avoided at any cost? I was using these for exactly the same reasons stated above. This year, I have had three seperate instances where the switch had to lose power (move, re-work pwr arrangements, etc), and all three times the PSU apparently gave up the ghost and refused to power back up. Nothing 'happened' funny power wise, not zapped or otherwise mistreated in any way. I think these units were of a vintage vulnerable to the bad chineese capacitor problem and I think whatever cap in the psu just went fizzle while it was operating, which would let the units continue running but once it lost power, would prevent a successful full power on start up. I was able to find and deploy the rps-675 (redundant power) after being burned this way three times, and it came in damm handy because there was a 4th event (another burned up 2970 psu) and this time the 675 kept it running till I was able to have an orderly replacement and maintinence window (with a 3560). I would reccomend deploying the rps units if you are going to use any cisco products with single power supply, but especially if you're going to be using the 2970's which have proven (in my shop) to be a (literally) dying breed. Thanks Mike, Well adding RPS, will add cost to final price. I plan to install these on low-density POP-sites where only 1 switch need, thus one RPS per switch. This is basically price of WS-C2960G-24TC-L or WS-C2960S-24TS-L new. So I need to think about before order these, thanks for sharing. We already tried several vendors(low-cost) for access switches to fill gap, but not really happy with them. ___ cisco-nsp mailing list cisco-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/