(usr-tc) configuring a second HiPer ARC
I'm installing a second HiPer ARC into a Total Control chassis and am having some difficulty integrating it into the system for automatic failover. I'm running into several issues. One problem I've run into is the IP address pools that the HARCs use. Is there a way for both HiPer ARCs to share the same IP address pool without stepping on each other? Or if one fails, can the other one automatically acquire the address pool of the failed HARC? Is there a way for the second HARC to be assigned the first HARC's IP address as a hot standby IP address? If neither HARC is configured as the owner of a modem card, how does the NMC determine (or is configured to specify) which HARC does route for a particular modem card? - To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
Also sprach Antonio Querubin
I'm installing a second HiPer ARC into a Total Control chassis and am having some difficulty integrating it into the system for automatic failover. I'm running into several issues.
One problem I've run into is the IP address pools that the HARCs use. Is there a way for both HiPer ARCs to share the same IP address pool without stepping on each other? Or if one fails, can the other one automatically acquire the address pool of the failed HARC?
Sort of. The best solution I've found is to set the Arcs to not pick IP addresses out of their pools by round robin and let them pick first available. Default is round robin. disable ip address_pool_round_robin Then you can setup two seperate pools on the Arcs for each to serve half the modems. Once those are set up, then for each Arc, you configure the *other* Arcs address pool in it. So, Arc A has address pool 1 then address pool 2 list, then Arc B has address pool 2 then address pool 1 listed. Follow? Since the pools are the same size, each Arc, in normal operation will only use the first pool, if the second Arc fails, then, when dynamic slot assignment rebalancing assigns the modems to the still good Arc, it will have the other pool of addresses, that the failed Arc was previously using, available to assign to the newly picked up modems.
Is there a way for the second HARC to be assigned the first HARC's IP address as a hot standby IP address?
I'm not aware of any way to do this, but with dynamic routing protocols, the benefit of this is limited.
If neither HARC is configured as the owner of a modem card, how does the NMC determine (or is configured to specify) which HARC does route for a particular modem card?
If you have dynamic slot assignment and dynamic slot assignment idle rebalancing enabled, then, given enough time, the modem cards will alternate between the Arcs for ownership. -- Jeff McAdams Email: jeffm@iglou.com Head Network Administrator Voice: (502) 966-3848 IgLou Internet Services (800) 436-4456 - To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
On Wed, 24 Jan 2001, Antonio Querubin wrote:
I'm installing a second HiPer ARC into a Total Control chassis and am having some difficulty integrating it into the system for automatic failover. I'm running into several issues.
It may not work at all. From the TC4.0 HiperArc release notes... MR 12602 Issue Dual HiPer ARC Fail-over not working. When using two HiPer ARCs, and one is pulled, the other fails to take ownership of the pulled ARCs HDMs. Status No Workaround currently exists -- Aaron Nabil - To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
It does work for me. I set it up so that one HARC owns all the cards and they are statically set. The backup is set to owner=no for all slots, but is dynamic. If the first HARC fails, the second one then comes in (not immediately, but it does come in) when the NMC notices the first one is gone and dynamically assigns all slots to the second one. Under normal operation, the second one is just sitting there, and I assign it the same IP pool as the primary HARC in the chassis. I've had this work in my test chassis, in a live controlled experiment where I had to shut down a chassis for other reasons so I pulled the primary HARC first to see if failover would work (it did; then I powered off the chassis), and once in an actual failure. That one was kind of rocky, though; the primary HARC failed, but tried several times to ressurect itself and take back control of the chassis. This didn't work, but was enough to disrupt the chassis several more times. Naturally, this happened while I was on vacation, and at a remote POP, not our Tokyo hub. By the time somebody got out there to remove the failed card, it had bounced up and down several times. It appears that failover will work fine in the even of complete failure, but if a HARC keeps coming up with "I'm not dead yet!" then your customers on that chassis can be in for a rough ride until someone dispatches it. Jonathan Byrne, CCNA Engineering Division <Large Hosting/Colo company with no public opinions> Tel: +81 3-5358-5291 Fax: +81 3-5358-5306 Aaron Nabil (nabil@SpiritOne.com) wrote:
It may not work at all. From the TC4.0 HiperArc release notes...
MR 12602 Issue Dual HiPer ARC Fail-over not working. When using two HiPer ARCs, and one is pulled, the other fails to take ownership of the pulled ARCs HDMs. Status No Workaround currently exists
-- Aaron Nabil
- To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
- To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
Running PRIs on hiperDSPs with a hiperarc, we have discovered a very very strange thing. It seems that if we "busy out" any of the channels on the PRI in TCM, the telco give the famous "The number you have reached, X X X - X X X X has been disconnected" ???? I've got limited data, I admit, because we naturally don't want this happening much. But anyone have any thoughts about this? I'm very surprised that there's anything we can do on our end that would make the telco give that message. Thanks in advance! David - To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
On Thu, 22 Feb 2001 drernst@kirkwood.hoosier.net wrote:
Running PRIs on hiperDSPs with a hiperarc, we have discovered a very very strange thing. It seems that if we "busy out" any of the channels on the PRI in TCM, the telco give the famous "The number you have reached, X X X - X X X X has been disconnected" ???? I've got limited data, I admit, because we naturally don't want this happening much. But anyone have any thoughts about this? I'm very surprised that there's anything we can do on our end that would make the telco give that message.
What's the NPANXX of your dial-in pool? -a - To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.
participants (6)
-
Aaron Nabil -
Antonio Querubin -
drernst＠kirkwood.hoosier.net -
Jeff Mcadams -
Jonathan Byrne -
nabil＠spiritone.com