Search this Blog

Monday, January 31, 2011

What would be better solution - ethernet 10gb uplink or fibre 8 gb between two switches?

Our company is going to make some changes in it's interior - so we are forced to make new rack room. Old one will stay as where it is, but we will connect two of rack rooms.Question is what two switches would be the best solution to connect two rack rooms? Our first opinion is there must be two separate ways of cables that connect those two switches (one in each rack room). Each cable must pass a lot of walls, and there is possibility of EM influence on cables.We want two separate ways so if one day something happens to one cable connecting two switches, we wouldn't lose connection between floors. What would be better solution - ethernet 10gb uplink or fibre 8 gb between those two switches? or maybe both if it is possible?And to make those two separate ways of cables redundant.

Tips:

  • Connect each rack rooms using fibre optic cables. Depending on the distance, you can either use multi-mode fibre (short haul/range) or single-mode (long haul/range). 10Gb uplink is possible. Here are your options:

o Layer 2 only switch:

o 2960S "D" series and 2360 (top-of-rack server switch)

o Layer 2/Layer 3

o 3560E/3560X and 3750X/3750E. If you have more finance then you can go with the 6500 (Sup32 or Sup720) or the 4500E (with Sup7).

  • Redundant ways are always nice .Make sure that they do not cross each other or run in parallel somewhere in the building. But would like to stress that depending on how you feel about your security between the two server rooms, there is a solution to the problem of someone ease dropping or connecting into the cables on the way, called 802.1ae. Link encryption at wire speed. The 3750x can do that, however it can not today do it towards another 3750x or 3560x, but that is something that is changing in the next release. That feature is perfect between server rooms. Also the 3750x has a nice power stack feature. and the "normal" stacking feature makes it a nice candidate also. If you need higher speeds fx for moving vmware servers or just data center capacity in and between the server rooms then I would look at the Nexus 5000 series
Citation - This blog post does not reflect original content from the author. Rather it summarizes content that are relevant to the topic from different sources in the web. The sources might include any online discussion boards, forums, websites and others.

Will there be a outage of the network cause of recalculating the Spanning Tree?

We have a simple question about Spanning Tree behavior (We use in our LAN the standard STP 802.1d).We want to add a new switch (STP enabled) to the LAN. This switch has a lower priority as the actual Root Bridge and will never become the new Root Bridge.

Question: If we add the new switch, will there be a outage of the network cause of recalculating the Spanning Tree, or will this only happen if the new Switch will become the Root Bridge?

Tips:

  • What do you mean that your switch has lower priority? It should be higher than the root bridge for example if root bridge has a priority configured like 16384 and your new switch has default which is 32767, you are all set. You need to cross check this on both the switches. If you add a new switch and it has a lower priority, taking the above example say lower than 16384, it will cause a STP recalculation over the entire network as this will become the new root bridge. If you have already placed the root bridge carefully on your network and this new switch has a higher priority, it will not cause the STP recalculation. The careful placement and configuration of the root bridge is the key in an STP enabled network.

  • The switch with the lowest bridge ID (BID) is elected as the root bridge. Bridge ID combination of bridge priority and mac address so both should be lower than all other switches to elect as a root switch. So your new switch should be higher than the existing root switch to not become as a root switch. If your new switch have same priority as existing root switch you can change it.And yes if the root bridge election process started to elect new root switch then there will be outage.
Citation - This blog post does not reflect original content from the author. Rather it summarizes content that are relevant to the topic from different sources in the web. The sources might include any online discussion boards, forums, websites and others.

How to upgrade IOS - 3750 Stack?

We want to upgrade the IOS on my stacked pair of 3750s. There is not enough flash to install the new IOS alongside the old one. Can we do this:

1. delete old IOS from flash1:

2. copy new IOS to flash1:

3. set system boot variable to use flash1:

4. reload and have both switches come up with the new IOS

5. check everything is OK

6. delete the old IOS from flash2:

7. copy new IOS to flash2:

We understand why it is not advisable to have mismatched IOS on the switches with a stack permanently. But is it OK for doing the upgrade, so that in case of problems we can change back to the old IOS more quickly?

We haven't found any documentation which explicitly states that it is possible to do this, and we don't have a way to test it ourself

Tips:

In order for switches in a stack to come up with the new image ,we need to have the image in the flash of all the member switches.

Therefore if you have the image in just one flash and not in other as you mentioned and you reload with the boot variable set the second switch will not come up with the new image.

Please refer to the following doc ,which will guide you to do the needful. Click here for the document . Another document can be found here .

Citation - This blog post does not reflect original content from the author. Rather it summarizes content that are relevant to the topic from different sources in the web. The sources might include any online discussion boards, forums, websites and others.

Thursday, January 27, 2011

IOS Upgrade on a 6500 with redundant SUPs in SSO mode

We’ve read the doc about how to upgrade the IOS on a Catalyst 6509 running redundat SUP720 in SSO mode. So basically the procedure is simple:

1. copy new IOS on disk0: and slavedisk0:

2. update boot variable

3. copy run start

4. hw-module module reset

5. redundancy force-switchover

But the question is: what is the downtime of the network/module when following these steps?

Or is there a better way to upgrade the IOS?

Tips:

If you are running SXI and later you can do a eFSU which is a hit less upgrade. Click here for further guideline.

But if you are running any other versions (SXF or SXH) the above mentioned procedure is correct but there will be downtime of 2 or more minutes because at one point both supervisors will be running a different image and will be RPR mode.Click here for the link to upgrade doc which you might already have seen.

Citation - This blog post does not reflect original content from the author. Rather it summarizes content that are relevant to the topic from different sources in the web. The sources might include any online discussion boards, forums, websites and others.
 
/* Google Analytics begin ----------------------------------------------- */ /* Google Analytics end ----------------------------------------------- */