9800 wireless controller High availability with the RMI and gateway reachability failover feature

today I'm gonna cover configuration of a che on 17 dot X code on the 1900 controllers I have previously covered how to configure a che on 16 dot X code however for those of you who are not aware 16 or ex corde has ax the gateway reachability check and the our my interface as well these are new features introduced in 17 dot X code okay so I would be covering those two features before I do that I'm going to clear the SI configuration on these controllers okay now once you notice one thing though not sure if you guys aware of this the way you clear the HEI configuration is you need to use clear chassis redundancy on the active controller okay the documentation clearly mentions that we need to run this command the standby controller which is the one on my left is going to reboot and it's going to come up with the blank configuration okay once it comes up with a blank configuration we obviously need to go in and configure it in order to get the GUI access once we access the GUI we will be dropped into day zero okay so I'll show you all this right now so I'm doing okay so immediately you can see the problem it's saying that RM is configured modification is not allowed ok so the first thing I'm going to do is I'm going to login into the GUI of the controller I will disable the H a from the GUI I'll see if I can do that looks like it says configuration applied successfully let me bring up the come yeah so there you go so looks like if you have RMI configured you cannot run this command clear chassis redundancy you have to do it from the GUI but as expected the standby controller has begun rebooting it's going to come up with pretty much a blank config will go in configure the IP addresses make sure we got GUI access okay also log into into the current active controller but but basically a redundancy should be disabled now that we have cleared the HT configuration so if I do show redundancy so chassis which it says chassis 2 is removed okay let's wait for this controller to come up in while I'm going to go back again into the active control though so as you can see here the redundancy has been disabled ok this is the current active control 196 basically this one ok so just to repeat if I do clear chassis redundancy from the GUI this time this time that controller will reboot it's going to come up with a blank and fake it's we have to basically assign the IP address and basically get the GUI access it's going to drop us into de 0 config will see that as well and parallely the current active controller is it going to have this redundancy configuration clear as well ok I'm just waiting for this controller to come up once this comes up I'll give the IP address make sure we've got cui access and then I'll come back okay guys so this controller has come up and I just wanted to quickly show you that this controller has been blanked out if we observe the hostname is now saying WL see if I do show IP interface brief there is no IP address is assigned anywhere in fact the service port which was up previously is now admin down and this controller basically has got no configuration okay so I'm just going to pause the video because as I said we need to arrange the GUI access so we can go in and configure the redundancy so I'll be right back in five minutes okay I'm back so I've configured the router sorry in that router I've configured the controller I think I forgot to bind the plus point just give me one moment let me just finish that but done so now I should have access to the controller okay this is 1902 I'm gonna login mmm I think I have not considered the authentication correctly in fact I think I missed configuring the authentication frankly so let's do that okay now here's what we're gonna do okay both the controllers I've got what they actually have got four cables okay okay before that as you can see here because we did the clear chassis redundancy not that I got the GUI access of the other controller the perform standby controller I'm now dropped into day zero configuration okay so I need to complete the day zero configuration so let's finish this country code is invalid select from available list okay I'm just gonna do it as US there's something wrong here because I cannot see the wireless management settings in the browser let's see yeah cool I'm going to leave all this some the waiters I have the service for IP address a wireless management right so what is management is gonna be actually this one be nine three three three three three I just let it configure all this but I know the the day zero configuration makes you do it again three is everyone next next asking for poisoning posture which is not required actually next and finish okay so it's gone ahead and configure a bunch of stuff which I'm gonna have to manually change okay so showing interface 10 gig 0 0 it actually configure citizen as a trunk port but it's not a trunk on the other side so let me fix this I wonder why we even have the day 0 consider it's real waste of time okay now I'll go in I'll explain the setup as well in just a minute mm okay this one just watch this forward mode access support access v9 3 I'd make sure I campaign 303 303 which is an IP address and I should open the gateway as well yeah it's reachable now so as I was saying I've got four cables okay four cables connected to both the controller's or sorry five cables okay one cable goes from the service ports one cable goes from the RP port wreden support that is one cable goes from ten K X zero zero one port one I call that second cable goes from port to port cable goes from port 2 and v cable goes from 4:3 okay for 1.2 1.3 all connected to a switch and those ports are on access ports port one is on VLAN to sorry for zeros and VLAN to port one is on B then 3 and port 2 is in return for okay so what I've done is I've configured the default gateway out VLAN to maybe see show IP route my default gateway is okay I think this is all happened because of the day 0 configuration let me remove the the default router it is added automatically my default route is out VLAN to my wireless management is on billion 3 though okay this is not set up I'm going to share out so as you can see a default gateway is pointing out to 2 to 1 which is v-nand to kind of like management VLAN but my wireless interface is just emitters VD on three okay so the first thing we're gonna do is we're going to go ahead and configure the redundancy so let's make sure that the day zero has completed okay so I'm locked into the glee of the second controller now when I configure the redundancy it's gonna ask you for a bunch of extra IP addresses which I'll show you in just a minute in case you you're coming from the $16 Xcode so in 16 or Xcode yeah there was no option to configure RMI if you're coming from air OS you're probably used to this though because in air OS you always had RMI so here this one unable see are mi plus RP I do have the option of going back to only RP okay so I think we all should be knowing this so this is similar to $16 Xcode where I put in my local IP so this is the IP address assigned to thee our we port okay but in this version we have the option of doing our mi plus RP which is what I'm going to be doing so I can show you the gateway reachability feature check now for the are my IP I'm gonna be using I mean let me just check one thing I can think for for for one okay from this one I can't because actually there's a problem with the cable I think on this controller so so what I'll do is because my gateways on VLAN to I will I will choose my are my IP from VLAN 3 which is the wireless VLAN okay so for chassis 1 I will select 3 3 3 for the chassis to also like 3 3 3 5 ok this will be automatically populated the it will start with the IP address 160 9.25 for the last two octet will be taken from here these two okay so local IP will become I'm not sure which chassis is this I think this is chassis one so local IP will become 169 254 3.4 and remote IP will be 169 254 3.5 okay let me show you that one here apply configuration applied successfully see exactly what I said right and the management gateway fellow what is automatically enabled you can disable this but it frankly it will be pointless if you are disabling it so again once I want source of confusion I'm now on the other controller I mean while I think this controller would have okay it's not a rebooted but one common source of confusion is do I need to reverse these entries in the second controller no you need to put the exact same IP addresses because read the text the text says rmii p4 chassis one here also it said the same thing the are my IP is already changed okay however the software is intelligent enough to know that this is chassis too so even though over here it sure local IP is 3.4 and remote as 3.5 this would be reverse here so here it's going to show me 169 254 3.5 in the remote IP is 169 254 3.4 let me show you that that's why again local IP should be 3.5 remote IP should be 3.4 yeah 3.5 and 3.4 okay it even has this little thing here saying reboot required okay so actually I don't know why they even say that because they may go ahead and reboot so as you can see here chassis 2 is reloading and the reason is stack merge which means it has detected that there is already a controller that's active in the network just going to reload and it's going to join the cluster ok once this joins we will just look at some for command outputs two of them are similar to what we had in 16 or Xcode the other two I think are mainly for the 17 dot X I'm just going to pause the video because this video is gonna take some time and I'll be right back ok so I just want to show you guys something so this controller booted up but it as you can see here it's stuck it says that chassis number 2 has it says that all chassis have been discovered accelerating discovery but it's stuck ok the reason for that is if I do show chassis command which is a very very important command okay you can see so this is chassis one ok this is chassis one you can see chassis 2 is removed actually that's because I've done a shot on the on one of the ports ok but if I scroll up you can see here the issue is V mismatch so V mismatch sans perversion mismatch and the reason for that is that chassis 1 which now the active chassis is running the 17.2 code which I mentioned earlier but the other chassis is actually come up with 17.1 code it does have the 17 tortue code on the flash so I'm going to go ahead once I get the access I'm going to go ahead and show you guys how to install the new code that's sitting there and flash through the CLI but now the immediate problem and let me restore things back the way they were so you can see the B mismatch error that I was talking about but the immediate problem is how do I get the access to the CLI of the of the node yeah because as you can see I'm I'm pressing enter nothing is happening this guy says the chassis is removed okay it says removed this one nothing is happening see same chassis - reloading reason lost both active and standby okay so that's wait for this chassis to reboot I think the it is basically established the communication back again because I I did a no shell on the RP port on my switch so now we're back again to that be mismatch message okay so we mismatch means version mismatch okay I think it's big and rebooting again okay so what I'm going to do is I'm going to go ahead and isolate that device from network by shutting down the switch ports so these ports are connected to 15 16 and 17 as I believe the other one is connected on 18 and 19 for sure and I believe there was one more port right so 15 16 17 let me check with what was it certain 1819 so shot nineteen I've shot eighteen okay twenty right so 20 is not working anyway so I'm not gonna bother but I think I think that's it I want to shut the RP board down so there should be no communication across the RP link so hopefully this controller should now come up as a active controller because it cannot detect the other controller once it does I'm going to go ahead and upgrade it okay I'm gonna install that 17.2 image and give it one more reboot and hopefully then they will form the H it together okay so let's just wait and see what happens okay so that looks to have done the trick the controller is now booting up and we should have CLI access shortly there you go so when quick come on to check the version in case you're not aware is show installed somebody okay try not to use the show version command this is a pretty good command see you can see here it shows the image is 17 1 1 as I said I do have the 17.2 code on flash in case you don't you can just copy it through either FTP TFTP I believe even HTTP is supporter now or you could even use a USB stick to copy the file and in this particular case I've already copied the file and I'm highlighting the file name this one so this is the method I use to upgrade any controller it has worked very well for me I do not use the GUI to upgrade my 9800 controllers I copied the file using any of these methods I have also copied the file through the GUI in case you're not aware that can be done I'll show you that later ok Pro once that's done I use this command install add file because it's on my blue flash I use blue flash in case you copy to the hard disk and use hard disk as well ok activate and commit ok that's it before you hit enter I mean it's gonna be is going to ask you whether you have changed the boot variable to packages or cons please make sure have done that otherwise the controller is going to put up with the old image again ok I'll show you what I mean I'll I'll show you on this controller so the boot variable when you use this method should be pointing to packages conf ok this ensures that they control the boots up in install mode ok there are two modes basically of installation on iOS XC devices one is the install mode which is preferred the other one is the bundle mode which is basically how it comes installed when you first buy the device okay so make sure the boot variable is set to packages don't conf do not set the boot variable to the actual binary file that's on flash okay so we have a message now saying that say this is the message I was talking about it's saying please confirm that you have changed the boot config to both flash packages or conf okay so if you haven't please hit no here go back to configuration mode and do boot system boot flash packages or cons in our particular case I've done it Sam Soon hell yes and now it's going to start the addition of the image along with the activation I'm just gonna pause the video because this does take a bit of time and it's pretty boring just sitting there watching this okay so I'll just pause the video and be right back just to show you the screen I got one more message here saying that this operation will require reader of the system do you want to proceed I'm going to type why it's activation is done I think it does a commit after this and it's going to reboot post that okay so as you can see here it's finished the comment okay now that is booting up let me go ahead and do the know shut on the interfaces so that the a chicken form smoothly I'm going to know shut the RP port first then I'm gonna know shut the other ports so the two ports okay so as you can see oh okay it looks like it's still booting up with the old image not sure what happened then but we'll take a look anyway you know okay you know what I think I do not set the boot very well okay that's fine we'll fix it so I'm gonna I'm going to no shut the boats okay so that this controller boots up and gives me the CLI access and then I'll come back show you that the boot variable is missing we will fix the boot variable and go through the upgrade process again so the controller has booted up as usual it has come up with the as expected let's come up with the old image 17.1 one as I said the reason for that is most likely the boot variable was missing so let's confirm that and it was the indeed the issue okay so I should have followed my own advice and I should have said the boot variable so what I'm gonna do is I'm going to go ahead and copy that boot variable from the from the other controller that's it then I'm gonna go ahead and use my command which is the [Music] install command I'm gonna search for my binary image which is this one I wanna say install add file boot flash like this activate activate okay that's because there is a space here by mistake map to activate and then commit okay that's it so I'm gonna let it run and do its thing since I've already shown you the the two places we need to press yes okay once this boots up I'll go ahead I mean before it boots up I will again no shirt the ports on the switch the RP port and the wireless management and the normal port 0 with like the v10 two interfaces connected by the default which points the default gateway okay and then I'll come back okay so the controller are now booted up and as you can see it's booted up with the correct image 17 dot X and we got a message here saying that waiting for remote chassis to join so it should join now so oops so showing be mismatched but I think that's probably a cosmetic thing yeah so all chassis is in this tag have been discovered accelerating discovery so that message should be gone now Stosh when we mismatch I think that's just something which should eventually go away let me just make sure that I'm running the correct version yep which I am so our chassis is in the stack have been discovered so as you can see here it has booted up with 17 - X version so I'm honestly not sure what issue is unless I have not shut some port properly I rather not know SATA ports so the the rear-engine supports are out okay that's confirmed I'm not sure what the issue is here oh you know what I want to do I'm just gonna go ahead and reload this box I think that was okay so if I don't know if you guys noticed but when I reloaded chassis one chassis two is reloading as well which means that there she was formed correctly I think that V mismatch was just something cosmetic I probably should have shown check the output of show redundancy but I don't use that command a lot so I completely forgot but anyway once these controllers boot up they should form the H a and we can test the gateway reachability feature okay so looks like the discovery is over chassis one looks to be booting up properly I'm not sure what's going on with chassis two though but let's see okay so Jessie one has indeed booted up and as we can see here chassis number two is showing ready okay so as you saw I didn't do anything I just rebooted the chassis and now both have come up and properly formed the HEA in fact chassis two is just rebooting it's just booting up as we speak so this is wait so the H async is in progress again take note installation mode is installed okay any time you use that command install add file activate commit it's going to install that image in install mode and the boot variable has to point to packages or cons do not point the boot variable to the actual binary image okay how do we check what happens in this particular scenario but because I've been just following my own rules blindly but I did face some issues earlier when I pointed to the binary image we can check that in the setup later but right now I would like to focus on DHA piece so I think this controller is almost up now what's the situation here it's still showing a sink in progress so one that's happening just to remind you VLAN three is my wireless management elin I do have an SP port which is not relevant here I have an RP port which is very much relevant my are my IPS are from the wireless management which is from VLAN three and in addition to that I've created another VLAN be 9/2 and my default gateway is pointing out wheel and - okay it's not pointing out the wireless management will just be round three by the way I think both should be ready now yeah there you go so the HEA has formed correctly now I have been using this command show chassis didn't come on actually showed redundancy which is also pretty ok you can use this command you can see your active slot 1 and slot 2 is standby heart okay it shows you the boot variable as well which is the package insert conf that we put in earlier and this is the two versions okay however both these commands do not show us the are my IP address so there is one new command called show chassis RMI if you run this command you will find you will get the re my IP address okay so three three three 0/24 is my wireless management VLAN and my are my IP addresses are from that VLAN okay first things first I'm gonna check if I can tame the rpi port IP address so that that's a common source of confusion for everyone so if you notice we cannot paying the IP address assigned to the ARP port okay in fact if you notice I cannot take my own IP address as well okay but we can think the RM IP so my Armagh IP is 3 3 3 4 I can think that I can also paying 3 3 3 5 which is the standby controllers are my IP address there you go my my own real IP is 3 3 2 3 2 which I can ping the the gateway is 3 3 3 1 which I can ping but I cannot paying 3 3 3 3 which is the real IP address of be scanned by controller ok this is the expected behavior ok in areas we could SSH into the or in my IP address so let's try that first I'm going to try and SSH into the active controllers out of my IP address so I've put in three three three four it's not working by the way so let me try in fact I should have tried three three three two first okay so this connection has timed out okay I suppose there is some issue with my configuration let me check that oh sorry I won't be able to SSH from my PC I can only access it from the switch so the stretch from the switch SSH - L I mean three three three the first thing I'm doing is I'm SS searching into the real IP address okay I'll show you that first mm-hmm so it's saying connection aborted can a ping it second thing let me try this is the standby controller IP address this is not going to work as you all know so I'm just trying to exit out from this connection okay let me try now the other my IP address it's also saying about it actually you know what I'll do let me go ahead and unable telnet access as well to the duty controllers because there might be some SSH incompatibilities which I don't want to end up trouble sharing the magic command is transporting put all let's try it again this time worried I'm gonna just tell MIT I want to tell me - the real IP address first okay yeah so definitely they were that was a message incompatibility see I'm on T controller now okay I'll try I'll try three it's not working that's the standby controller okay once this times out I will try the RM IP address of primary controller I think we should be able to access the primary controller through both the real IP address as well as the arm IP address if you remember in air OS we can access the standby controller using the our my IP so I'm gonna check if it works here as well just waiting for this to timeout okay so there you go I am able to access the primary controller see this is the exact same controller so chassis chassis one and I'm going to try to access the standby and I can so the behavior is similar to air OS see I'm on this time by controller by the way this is you can't type anything because the standby console is disabled okay I don't think we can't even read over can we no we can't okay can we ping and check things No okay cool oh okay here's a new one so now that I've generated I can't exit okay that's fine um I don't really care about exiting out from here and this would eventually time or anyway let's go back to our our actual scenario right so here's what I'm going to do I'm going to break the communication between the primary controller and the Gateway oh I have a couple of ways of doing that I could either change the VLAN membership which I have tested and it works fine so what I'm going to do this time is I'm going to just shut the interface down completely okay let's see if that triggers available so switch I've just shot the interface down on the switch which means he should now lose the ability to the Gateway let's see they perish YP route C there is no gateway okay actually this is what happened last time as well okay let me do one thing okay let me change the way that membership and check it once because this can't be a coincidence okay Gateway is back now can I ping the Gateway I should be able to ping the Gateway yeah I can think so I'm gonna change this time I'm not gonna shut the port I'm going to change the VLAN membership of the of the port okay nothing is happening it does not take this much time guys so what I suspect what I suspect is the gateway has to be reachable on DM on the wireless management it can't be out on any interface okay let me change the VLAN membership of the wireless management interface and see what happens in fact I just shut the port down so I've shut the wireless management port on the on the ActiveX control see the three three three Network hell disappeared okay yes this is not making any difference so I guess now we got the working situation the working situation is very very simple I'm just going to go ahead and change it back to wireless management interface VLAN - okay it's done and I think they should work now let's see I'll shut the port first since that's an easy way of triggering of failure so my wireless management is VLAN - my gateway is pointing out we land - and I've shut the physical port physically interface on the switch if this doesn't trigger available I don't know what will okay this is not supposed to take so much time so let me try one thing and the pinoche a deport and I'll change the VLAN membership okay okay I've changed the wheel and membership something wrong because I stood this before well you know what maybe the energy management okay but I think everything has to match okay I'm gonna fix this issue again one second I think let me try let me just do what I did last time okay so wireless VLAN three that's it a new IP route okay see you then I'm gonna come here 3.0 3.1 okay this is done and I'm gonna shut the port down on the so I've shut the physical interface on the switch okay so everything is matching now my autumn is on VLAN three yeah there you go see it's reloading now so basically if the condition is my wireless management my autumn I am a default gateway has to be in the same VLAN okay only then it works see we have now taken over and the reason for the taking or should be gateway something gateway not reachable I remember as I've seen this so redundancy switch over history so active lost gateway okay current active - Jessie - that's me previous active is chassis one that's currently on the right and it's rebooting okay so this is the gateway reachability feature check one last thing I want to cover is so gateway reachability feature is again an optional feature of the redundancy but under the useful feature of the RMA is if the RP if the RP port goes down the controller will not land will not land instead brain scenario okay I'll show you what I mean now before that I'm gonna show you one more thing what's the standby controller comes up it's going to be in the state called standby recovery okay let me show you that okay so the controller is booting up okay I'll show you that it'll be in standby recovery okay we will then do a no shot on the porch so that the so that it's start to normal and then I want to do one more test before I show you the before I shot the RP port down and show you something called as active recovery okay okay so it's up now if you do show chassis that shows ready so redundancy doctor standby standby out okay I thought yeah there you go I was just about to say something is not right so see it's now in recovery mode okay this is called a standby recovery let's check the outputs again so slot two is me active you can see here as well loss switch over reason is active lost gateway and for some reason it's still showing slot 1 as standby hot let's check so chassis still shine ready ready but you can clearly see it in recovery mode now let me recover this condition first so yeah while we are in this recovery mode I want to do this one more test okay before we I'm going to go in the controller I'm gonna do one small config change I'll create maybe 1 billion because I want to see if they replicate okay what about a villain so I'm gonna add one I'm gonna add 1 billion here I call it B then 10 name is test test villain okay apply to device invalid VLAN name okay I guess it likes underscores or something okay I'm gonna save this configuration let's see if it any change here No and showing active and standby heart okay go ahead and fix this condition so no shirt on this on the switchboard okay I suspect the I suspect least analyzable to reboot now okay so I I think the one on my right is going to reboot let's see because when this condition gets fixed I want to see if the configuration gets replicated or not oh sure what's going on here okay something I've done I've done right here wait 15:18 okay I have not unshot the correct port guys it's my mistake so let me do that no shut okay they should fix this issue by the way what it's still showing ready ready okay so I guess there's no way to really know so you see like if you usually customers will not have consoles connected maybe to the welc so if you look at if you're on the shape air and you see stand by heart we can't really draw a conclusion from that what's happening all IP the pinging so not sure what the issuance okay there you go I think it's just slow so you can see here right now it's showing standby console disabled and the recovery message is gone here's what I'm gonna do I'm gonna do a redundancy switch over okay for switch over so this the current active is going to reboot and this one's going to take over it as the standby I'm sorry is the active yep so Jessie one is now active how do I know that run show chassis it's showing one is active I'm gonna log in again and see if my VLAN was changed I mean if it added a new villain which is VLAN 10 because suspect fields see a lot of cases where customer things the standby was done by heart just few changes well unless it has the change then that's fine I don't really care if it has to change okay it does okay I guess that's fine then okay I'm just gonna pause the video and wait for this guy to come up okay I just thought I'd shared this with you so this guide you'll reboot but if you notice it did discover but it says something like non-participant detector and some critical process has failed so it's basically rebooting again I'm not sure why what's the reason for that it says something like it's even generated a co report by the way we have a system called report at core this thing we might take a look at that later I'll pull it off from the controller but let's not worry about that now let's just proceed with the essay okay home so let's see what happens now okay so all chassés have been discovered accelerating discovery I think it should be adjacent in progress now okay it's showing me as initializing on the active controller okay so it looks okay now I think it began booting yep and we can see it a sink in progress what should happen last time stirring accessing in progress so I'm gonna wait for this to become their turn to ready the next time I want to do the next test that we're going to do is I'm going to break the connectivity between standby controller and the Gateway okay I repeat between standby controller and the Gateway last time I broke the connectivity between active control and the Gateway and we noticed that active control the rebooted and came up as came up in recovery mode as a standby correct but this time I'm going to break the connectivity between the standby controller and the Gateway and we're going to see if that changes anything so I just want to see if this time by goes into recovery mode okay if it doesn't then I think we have a problem because if I do a failover now manually nothing's gonna work right which is what I'm trying to test okay so it looks like we are still showing why is it showing edge a sink okay it's ready now so that's it this time I do not shut the interface by the way I just changed the VLAN of the switch port that should be good enough so there you go so my standby controller is in recovery mode okay currently again I'm a bit water that I don't see any change here's what I'm gonna do now I'm gonna do one small change here I'm gonna add one bill and I'm going to call it B down 15 15 apply save configuration' okay and now I think all the hell is gonna break loose I'm gonna do a redundancy switchover force let's see the T what happens just one is reloading oh my god so this guy's active so we are in deep trouble now because I can't pick my gateway for sure that's precisely why I was in standby recovery correct and the worst part now is if I go log into the GUI I probably will not have my VLAN I'm honestly not liking the situation it's a bit worrisome from a customer point of view okay so I do have a VLAN but you know what it should not let me it should not let me failover I'm just gonna see what happens once this guy puts backup because remember he does have the reach ability to the gateway so logically he should just grab the membership the the ownership of the the active redundancy so let's see okay so chassis one has all chassés have been discovered so it's initializing it's just booting up now so it should be Ajay Singh I think is just going to come up and sit like a standby controller well I'm hoping and I can just hope that once this guy fully boots up and moves into ready this is going to be another failover because the current active controller doesn't have reach ability to the gateway correct so I'm hoping they got the illogic right let's see okay so looks like it's come up as standby both are ready ready now is the moment of truth as we say the current active controller doesn't have visibility to the Gateway and we have a running 17 dot X code with our mi plus RP enabled and the gateway check is also enabled we just wait for a few minutes okay good so this guy I guess he realized that and he is rebooted and this guy's becomes active device perfect so I guess I'm happy now okay so I just fixed the VLAN membership on the switchboard so when this guy boots up he should boot up as a normal act as a normal standby controller okay so the last topic of the day is undercover is in the earlier code in 16 dot X code the only way for the two controllers to know if they were both alive was through the RP port okay there have been situations where in the RP connectivity is disrupted or has been broken and the both the controllers have landed up in active active this is not a very desirable situation as it's going to all the hell is going to break loose since the access points are going to disassociate with the controller from the controller and unless you have a and you probably will not even be able to access the right controller because both are going to ARP for the virtual IP right so this is called split brain so using this our ml feature we can avoid this problem because even if the RP connectivity is broken the two controllers will check over the RMI port or they are my interface I mean if they're able to reach each other okay if they can't reach over our mi interface I believe the active controller takes the state of active recovery it goes into a recovery mode basically and it sits there like a standby which is the funny part the other controller takes over as the active control though and that's exactly what we're going to check now I'm just waiting for this controller to boot up and for redundancy to be restored and then quickly do our test before that I'll just make sure that we have the what audience so basically there is no configuration loss so it's saying all chassis is have been discovered okay so I guess we are in good shape now it's initializing so we should be okay enough so configuration we then I should see 10 and 15 if I do I'm presuming everything is good okay ah so just as a reminder the controller on the right is the active controller its chassis one the control on my left is a standby controller and its chassis - I'm gonna break the connectivity between the two controllers only on the RP port mind you okay so this boot it up so it still having a je sink in progress okay so he's currently in standby I should be good okay now um I'm gonna break the RP connectivity as I said okay that's done see immediately this guy has become active controller it's showing removed this guy will show the other guys removed okay see I wonder let me check one thing on the switch so if I think so I am able to ping the the IP I can think the physical IP address but I'm not able to paint the out of my IP address okay there you go so the the former active I mean how should I say it the the the I mean the current active sorry has moved into something called the recovery mode and this guy is completely isolated as well okay let's see if I can ping the gateway which I can actually this is very interesting scenario so I'm gonna do a show at the ARP VLAN 3 so I can see 3 3 2 3 2 it's pointing to this guy can i ping okay so I don't know if you can see this but the active controller I it did not reboot but it became and went into recovery mode and the standby controller has now become active okay see I don't think the gateway in fact even even when I check from the switch if I check the ARP entry for 3 3 3 2 which is the wireless management IP address it basically points to chassis to basically points to this Mac ok I can ping 3 3 3 5 4 3 3 5 4 2 2 3 5 ok that's the other my IP address of this guy I think let's check sorry wrong command so I won chassis to my IP address autumn is 3 3 3 5 ok so I can think they are my IP address I can think the management IP address which is 3 3 3 2 I can't bring 3 3 3 3 and I can't ping 3 3 3 4 as well ok which makes perfect sense so last lastly I'm gonna I'm just going to go in and do do one change ok just for the sake of variety I'll configure one parameter map and when ACL ok and I'll check out that country got replicated so this is where you go and configure ACL by the way I'm going to call this test it's going to be an external ACL sequence 1 permit IP any-any is very simple easier I'm also gonna go and configure overbought parameter map which is from here I'm going to call it test uploaded device and I'm just gonna give it some some dummy URL okay like this this all doesn't matter by the way and lastly I'm going to change the virtual IP address to for our for our photo photo for our for I know yeah had a lot of changes I know so ok done now I will go and recover the situation on the ok on the switch I mean yep so that's done situation is recovered let's see still showing removed it's eventually going to realize yeah see so it started reloading saying stack much and I know this is a corner case for what I do is exactly the core file generated okay what I'll do is give me one second I'm just gonna do something weird I'm gonna shut the I'm gonna shut the interfaces I mean let's say interfaces I mean the are people it's another thing because I want to check one thing I want to check whether the change I did was it replicated or not okay so I have shut all the interfaces on this controller let me confirm that so you can see nothing is connected okay this guy should come up as the active controller and then you will login to the GUI and check whether the changes were pushed or not when it was in recovery mode I'm just gonna pause the video and be right back okay um looks like it's just putting up so it's booted up now okay obviously the other device is completely isolated so it's not even showing up here let's go back and check the configuration so I did a couple of changes I added an ECL added a parameter map where both okay and I also change the virtual IP address let's go and check each of those things one by one and one that's happening before I forget I'm gonna go ahead and do the know shot as well for the interfaces I'll go do that in some time so okay so we see the parameter map is gone I mean it's not there rather is here is out there as well and let's check the what should I P address was that I changed that to four four four four it's still one integer zero to that one okay so this is a real problem for us if customer hits this situation the I'm sure it's a corner case but it's a very valid situation to be in okay so I brought up the interfaces on this device okay I didn't bring up the RP port interface bring that up as well okay let's check now you can forget about port number three okay so all that up showing what recovery recovery mode why now it's not the recovery mode because I can think gave me yep it's rebooting now I'm going to pause the video while the reboots

Channel: Yogesh Gawankar

Views: 2,503

Keywords:

Id: 1d3642D57Nc

Length: 85min 54sec (5154 seconds)

Published: Sun May 17 2020