Re: [Veritas-ha] LLT heartbeat redundancy
This is not a recommended practice under Windows. However, if NIC Teaming underneath can provide 1 MAC address and device to LLT then this would work. Hasn't been tested and is not supported by Symantec though. -Original Message- From: andreas.lundg...@steria.se [mailto:andreas.lundg...@steria.se] Sent: Wednesday, May 13, 2009 9:39 AM To: Sandeep Agarwal (MTV) Cc: im...@inter.net.il; veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy Since I seem to be about the only windows user of vcs in the world ;-) I wonder a bit about this - I was under the impression this was not recommended practice in a windows VCS environment, but it would be supported according to these mails? /a Sent from a p990i Reply Header Subject: Re: [Veritas-ha] LLT heartbeat redundancy Author: "Sandeep Agarwal (MTV)" Date: 2009 May 13th 18:23 LLT does support Layer 2 link aggregation and can work over trunks/bonds/aggregated links as long as a single device is presented to it: Linux - Bonding Solaris - Sun Trunking (now link aggregation/dladm in Solaris 10) HPUX - Auto Port Aggregation AIX - Etherchannel IPMP is at Layer 3 and Bonding and the others are at Layer 2 - hence we can't compare the two. It would be better to compare Sun Trunking and Linux Bonding. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Wednesday, May 13, 2009 2:38 AM To: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Wednesday 13 May 2009 06:04:38 John Cronin wrote: > Getting slightly off topic, but still somewhat relevant. > > Linux has many flavors of Ethernet bonding. To be sure, link > aggregation resulting in increased bandwidth is generally supported on a single switch. > However, Linux does have an active-passive bonding that is > specifically intended for HA solutions. AIX has a similar > configuration with the unfortunate name of EtherChannel Network Backup > Interface - it does NOT rely on Cisco EtherChannel to work. Both of these create a "virtual NIC" > that hides the complexity, making the interface group appear to be a > single NIC. You don't need a bunch of switch link aggregation magic > (802.11ad or > EtherChannel) to implement active-passive NIC failover in this manner. > > In my experience, both Linux and AIX Ethernet bonding are easier to > use than Sun IPMP, and they also are far more reliable. I have a lot > of experience with all three of these, and in my opinion IPMP is the > worst - I have experienced many "false failures" with IPMP, and I have > had to do a bunch of silliness with static routes to make it work in > certain environments (prior to the new link based IPMP - but it has > issues of its own too). I wish Sun wou This email originates from Steria AB, Box 544, SE-182 15 Danderyd, Sweden, +46 8 622 42 00, http://www.steria.se. This email and any attachments may contain confidential/intellectual property/copyright information and is only for the use of the addressee(s). You are prohibited from copying, forwarding, disclosing, saving or otherwise using it in any way if you are not the addressee(s) or responsible for delivery. If you receive this email by mistake, please advise the sender and cancel it immediately. Steria may monitor the content of emails within its network to ensure compliance with its policies and procedures. Any email is susceptible to alteration and its integrity cannot be assured. Steria shall not be liable if the message is altered, modified, falsified, or even edited. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
Since I seem to be about the only windows user of vcs in the world ;-) I wonder a bit about this - I was under the impression this was not recommended practice in a windows VCS environment, but it would be supported according to these mails? /a Sent from a p990i Reply Header Subject:Re: [Veritas-ha] LLT heartbeat redundancy Author: "Sandeep Agarwal (MTV)" Date: 2009 May 13th 18:23 LLT does support Layer 2 link aggregation and can work over trunks/bonds/aggregated links as long as a single device is presented to it: Linux - Bonding Solaris - Sun Trunking (now link aggregation/dladm in Solaris 10) HPUX - Auto Port Aggregation AIX - Etherchannel IPMP is at Layer 3 and Bonding and the others are at Layer 2 - hence we can't compare the two. It would be better to compare Sun Trunking and Linux Bonding. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Wednesday, May 13, 2009 2:38 AM To: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Wednesday 13 May 2009 06:04:38 John Cronin wrote: > Getting slightly off topic, but still somewhat relevant. > > Linux has many flavors of Ethernet bonding. To be sure, link > aggregation resulting in increased bandwidth is generally supported on a single switch. > However, Linux does have an active-passive bonding that is > specifically intended for HA solutions. AIX has a similar > configuration with the unfortunate name of EtherChannel Network Backup > Interface - it does NOT rely on Cisco EtherChannel to work. Both of these create a "virtual NIC" > that hides the complexity, making the interface group appear to be a > single NIC. You don't need a bunch of switch link aggregation magic > (802.11ad or > EtherChannel) to implement active-passive NIC failover in this manner. > > In my experience, both Linux and AIX Ethernet bonding are easier to > use than Sun IPMP, and they also are far more reliable. I have a lot > of experience with all three of these, and in my opinion IPMP is the > worst - I have experienced many "false failures" with IPMP, and I have > had to do a bunch of silliness with static routes to make it work in > certain environments (prior to the new link based IPMP - but it has > issues of its own too). I wish Sun wou This email originates from Steria AB, Box 544, SE-182 15 Danderyd, Sweden, +46 8 622 42 00, http://www.steria.se. This email and any attachments may contain confidential/intellectual property/copyright information and is only for the use of the addressee(s). You are prohibited from copying, forwarding, disclosing, saving or otherwise using it in any way if you are not the addressee(s) or responsible for delivery. If you receive this email by mistake, please advise the sender and cancel it immediately. Steria may monitor the content of emails within its network to ensure compliance with its policies and procedures. Any email is susceptible to alteration and its integrity cannot be assured. Steria shall not be liable if the message is altered, modified, falsified, or even edited. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
LLT does support Layer 2 link aggregation and can work over trunks/bonds/aggregated links as long as a single device is presented to it: Linux - Bonding Solaris - Sun Trunking (now link aggregation/dladm in Solaris 10) HPUX - Auto Port Aggregation AIX - Etherchannel IPMP is at Layer 3 and Bonding and the others are at Layer 2 - hence we can't compare the two. It would be better to compare Sun Trunking and Linux Bonding. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Wednesday, May 13, 2009 2:38 AM To: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Wednesday 13 May 2009 06:04:38 John Cronin wrote: > Getting slightly off topic, but still somewhat relevant. > > Linux has many flavors of Ethernet bonding. To be sure, link > aggregation resulting in increased bandwidth is generally supported on a > single switch. > However, Linux does have an active-passive bonding that is > specifically intended for HA solutions. AIX has a similar > configuration with the unfortunate name of EtherChannel Network Backup > Interface - it does NOT rely on Cisco EtherChannel to work. Both of these > create a "virtual NIC" > that hides the complexity, making the interface group appear to be a > single NIC. You don't need a bunch of switch link aggregation magic > (802.11ad or > EtherChannel) to implement active-passive NIC failover in this manner. > > In my experience, both Linux and AIX Ethernet bonding are easier to > use than Sun IPMP, and they also are far more reliable. I have a lot > of experience with all three of these, and in my opinion IPMP is the > worst - I have experienced many "false failures" with IPMP, and I have > had to do a bunch of silliness with static routes to make it work in > certain environments (prior to the new link based IPMP - but it has > issues of its own too). I wish Sun would add an active-passive > capability to their new link aggregation capability (dladm) that works > across switches. If they did that, they would have the same > capabilities as Linux and AIX network bonding, with similar ease of > use. It should be fairly trivial to implement. > > The one advantage that IPMP has in active-active mode (e.g. NOT link > based) is that it can detect IP connectivity issues (via ping - not > just Ethernet link detection) on all NICs in an interface group. > However, it is usually issues with the IP connectivity checking that > cause all my problems with IPMP, and I would gladly trade it for a > simple link based virtual solution that looks like a single link to me. The linux bonding driver can do ARP test, in order to detect uplink failures. This is not a layer 3 check, but in most solutions and configurations, it is a suitbale replacment. > > I have never used Linux Ethernet bonding or AIX Etherchannel Network > Backup Interface for VCS heartbeats, but I am pretty certain they > would both work fine. That said, I am not sure they would provide any > significant benefit over a "traditional" VCS heartbeat network > configuration, using the same number of "real" NICs. The benefit is that with such bonding method, you can survive the failure scenario I've described in my first email :) It is a fact Symantec understands that, as they are trying to solve it internally in LLT :) > > -- > John Cronin ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Wednesday 13 May 2009 06:04:38 John Cronin wrote: > Getting slightly off topic, but still somewhat relevant. > > Linux has many flavors of Ethernet bonding. To be sure, link aggregation > resulting in increased bandwidth is generally supported on a single switch. > However, Linux does have an active-passive bonding that is specifically > intended for HA solutions. AIX has a similar configuration with the > unfortunate name of EtherChannel Network Backup Interface - it does NOT > rely on Cisco EtherChannel to work. Both of these create a "virtual NIC" > that hides the complexity, making the interface group appear to be a single > NIC. You don't need a bunch of switch link aggregation magic (802.11ad or > EtherChannel) to implement active-passive NIC failover in this manner. > > In my experience, both Linux and AIX Ethernet bonding are easier to use > than Sun IPMP, and they also are far more reliable. I have a lot of > experience with all three of these, and in my opinion IPMP is the worst - I > have experienced many "false failures" with IPMP, and I have had to do a > bunch of silliness with static routes to make it work in certain > environments (prior to the new link based IPMP - but it has issues of its > own too). I wish Sun would add an active-passive capability to their new > link aggregation capability (dladm) that works across switches. If they > did that, they would have the same capabilities as Linux and AIX network > bonding, with similar ease of use. It should be fairly trivial to > implement. > > The one advantage that IPMP has in active-active mode (e.g. NOT link based) > is that it can detect IP connectivity issues (via ping - not just Ethernet > link detection) on all NICs in an interface group. However, it is usually > issues with the IP connectivity checking that cause all my problems with > IPMP, and I would gladly trade it for a simple link based virtual solution > that looks like a single link to me. The linux bonding driver can do ARP test, in order to detect uplink failures. This is not a layer 3 check, but in most solutions and configurations, it is a suitbale replacment. > > I have never used Linux Ethernet bonding or AIX Etherchannel Network Backup > Interface for VCS heartbeats, but I am pretty certain they would both work > fine. That said, I am not sure they would provide any significant benefit > over a "traditional" VCS heartbeat network configuration, using the same > number of "real" NICs. The benefit is that with such bonding method, you can survive the failure scenario I've described in my first email :) It is a fact Symantec understands that, as they are trying to solve it internally in LLT :) > > -- > John Cronin ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Thursday 07 May 2009 14:11:02 John Cronin wrote: > On thing to consider - don't you want to know when you lose a heartbeat > link? I know the lost link will be noted in /var/log/messages or wherever > your syslog configuration indicates, but I would want it in my VCS logs > somewhere too. Of course - But this messages appears during normal operations - i.e. when the two links are active and working. Veritas says that when the heartbeat links are not isolated, this is the result. This whole thread is about the new feature that allows HB links to be not-isolated, in order to allow full mesh. > > Also, are you using one or two bonded pairs? In other words, do you still > have two private heartbeats? If you are using a single pair, are you also > using a low-priority heartbeat on your "real" network? We are talking about 2 physical connections, not bonded, and the low-priority link is irrelevant. If you will read my original question, you'll understand the problem in the current design of the HB links. > > Also, are you using I/O fencing to protect your data? No, I/O fencing is not supported for Veritas Cluster File System for Oracle RAC, but this is also irrelevant, please read my original question. > > The reason I ask is this - what is going to happen if/when you lose one > link, then a few days later you lose the other link? The answer is that > BOTH of your servers will decide they are the only survivors, and they will > BOTH attempt to bring your storage online and start the application. If > you don't have I/O fencing enabled, this will most likely lead to data > corruption - time to start looking for the backups. > > By default, in the "normal" VCS heartbeat configuration, if you lose one > heartbeat link, and then more than 16 seconds elapse, you go into a > "jeopardy" state. If you then lose the other heartbeat link, the service > groups are "Autodisabled", to prevent a network induced split brain > condition from causing data corruption. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Wednesday 13 May 2009 05:27:00 Hudes, Dana wrote: > I don't care what tricks Linux plays or what they call it. From a > network perspective, true bonding requires connection to the same > switch/router and is done at the link layer. You don't have 2 IP > interfaces, you have one. The bits go out round-robin. It requires > support by the link partner/peer (i.e., you could do it with multiple > crossover connections between two hosts which have the appropriate > drivers). No, you have 2 interfaces, and the IP binding jumps from one interface to the other. This is very much different than how other modern operating systems works (FreeBSD, Linux etc.) - They present a logical interface which hides the underlying interfaces, and therefore let you configure a consistent name in the application which requires to know the name of the interface (LLT, for example). > > Solaris IPMP supports both active/active and active/passive (at least as > of Solaris 10). > > http://docs.sun.com/app/docs/doc/816-4554/eobra?l=en&a=view&q=ipmp I know how IPMP works, and we use it quite a lot, but IPMP cannot help me solve the failure scenario I've described in my first post. > > Note that you can configure multiple VLANs into an IPMP group on one > interface. This doesn't give you any failover/failback capability: if > the link peer goes away, your link is down. > > With active/active, you can have IPMP load spreading. This is only > effective with multiple IP destinations (even if those multiple IP > destinations are logical interfaces on one NIC on one host). > > IPMP groups are strictly the same media type but you can (and I have) > have a copper 100baseT port in the same group as a 1000baseF port: > they're both Ethernet. This does not work to backup an ATM interface > with an Ethernet interface etc. (I'm not sure if you could backup a LANE > interface with a classical IP-over-ATM interface and I haven't any ATM > switches to try it with these days). > > Since this is an IP-level redundancy and LLT doesn't use IP it's not > going to help VCS. > > IPMP is a replacement / successor to Sun Enterprise Alternate Pathing > (and is incompatible with it). In short, AP was for the Enterprise 1 > only and provided a means of rerouting the connection to an HBA (whether > for storage or network). It is replaced by IPMP and MPXIO (which is > better than Veritas DMP but only in Solaris 10+). > > Bonding isn't really something I expect to see in an Ethernet > environment but perhaps that's because I used to do it in an ATM > environment years ago. I'll have to look into what modern Ethernet LANs > do in this regard. Bonding (read - failover link aggregation) really helps HA solutions, and from "smarter" implementations of said feature, LLT can benefit a lot. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
Getting slightly off topic, but still somewhat relevant. Linux has many flavors of Ethernet bonding. To be sure, link aggregation resulting in increased bandwidth is generally supported on a single switch. However, Linux does have an active-passive bonding that is specifically intended for HA solutions. AIX has a similar configuration with the unfortunate name of EtherChannel Network Backup Interface - it does NOT rely on Cisco EtherChannel to work. Both of these create a "virtual NIC" that hides the complexity, making the interface group appear to be a single NIC. You don't need a bunch of switch link aggregation magic (802.11ad or EtherChannel) to implement active-passive NIC failover in this manner. In my experience, both Linux and AIX Ethernet bonding are easier to use than Sun IPMP, and they also are far more reliable. I have a lot of experience with all three of these, and in my opinion IPMP is the worst - I have experienced many "false failures" with IPMP, and I have had to do a bunch of silliness with static routes to make it work in certain environments (prior to the new link based IPMP - but it has issues of its own too). I wish Sun would add an active-passive capability to their new link aggregation capability (dladm) that works across switches. If they did that, they would have the same capabilities as Linux and AIX network bonding, with similar ease of use. It should be fairly trivial to implement. The one advantage that IPMP has in active-active mode (e.g. NOT link based) is that it can detect IP connectivity issues (via ping - not just Ethernet link detection) on all NICs in an interface group. However, it is usually issues with the IP connectivity checking that cause all my problems with IPMP, and I would gladly trade it for a simple link based virtual solution that looks like a single link to me. I have never used Linux Ethernet bonding or AIX Etherchannel Network Backup Interface for VCS heartbeats, but I am pretty certain they would both work fine. That said, I am not sure they would provide any significant benefit over a "traditional" VCS heartbeat network configuration, using the same number of "real" NICs. -- John Cronin On Tue, May 12, 2009 at 10:27 PM, Hudes, Dana wrote: > I don't care what tricks Linux plays or what they call it. From a > network perspective, true bonding requires connection to the same > switch/router and is done at the link layer. You don't have 2 IP > interfaces, you have one. The bits go out round-robin. It requires > support by the link partner/peer (i.e., you could do it with multiple > crossover connections between two hosts which have the appropriate > drivers). > > Solaris IPMP supports both active/active and active/passive (at least as > of Solaris 10). > > http://docs.sun.com/app/docs/doc/816-4554/eobra?l=en&a=view&q=ipmp > > Note that you can configure multiple VLANs into an IPMP group on one > interface. This doesn't give you any failover/failback capability: if > the link peer goes away, your link is down. > > With active/active, you can have IPMP load spreading. This is only > effective with multiple IP destinations (even if those multiple IP > destinations are logical interfaces on one NIC on one host). > > IPMP groups are strictly the same media type but you can (and I have) > have a copper 100baseT port in the same group as a 1000baseF port: > they're both Ethernet. This does not work to backup an ATM interface > with an Ethernet interface etc. (I'm not sure if you could backup a LANE > interface with a classical IP-over-ATM interface and I haven't any ATM > switches to try it with these days). > > Since this is an IP-level redundancy and LLT doesn't use IP it's not > going to help VCS. > > IPMP is a replacement / successor to Sun Enterprise Alternate Pathing > (and is incompatible with it). In short, AP was for the Enterprise 1 > only and provided a means of rerouting the connection to an HBA (whether > for storage or network). It is replaced by IPMP and MPXIO (which is > better than Veritas DMP but only in Solaris 10+). > > Bonding isn't really something I expect to see in an Ethernet > environment but perhaps that's because I used to do it in an ATM > environment years ago. I'll have to look into what modern Ethernet LANs > do in this regard. > > > = > Dana Hudes > UNIX and Imaging group > NYC-HRA MIS > +1 718 510 8586 > Nextel: 172*26*16684 > = > > -Original Message- > From: Imri Zvik [mailto:im...@inter.net.il] > Sent: Tuesday, May 12, 2009 9:54 PM > To: Hudes, Dana; veritas-ha@mailman.eng.auburn.edu > Subject: RE: [Veritas-ha] LLT heartbeat redundancy > > Hi Dana, > > We were talking about L
Re: [Veritas-ha] LLT heartbeat redundancy
I don't care what tricks Linux plays or what they call it. From a network perspective, true bonding requires connection to the same switch/router and is done at the link layer. You don't have 2 IP interfaces, you have one. The bits go out round-robin. It requires support by the link partner/peer (i.e., you could do it with multiple crossover connections between two hosts which have the appropriate drivers). Solaris IPMP supports both active/active and active/passive (at least as of Solaris 10). http://docs.sun.com/app/docs/doc/816-4554/eobra?l=en&a=view&q=ipmp Note that you can configure multiple VLANs into an IPMP group on one interface. This doesn't give you any failover/failback capability: if the link peer goes away, your link is down. With active/active, you can have IPMP load spreading. This is only effective with multiple IP destinations (even if those multiple IP destinations are logical interfaces on one NIC on one host). IPMP groups are strictly the same media type but you can (and I have) have a copper 100baseT port in the same group as a 1000baseF port: they're both Ethernet. This does not work to backup an ATM interface with an Ethernet interface etc. (I'm not sure if you could backup a LANE interface with a classical IP-over-ATM interface and I haven't any ATM switches to try it with these days). Since this is an IP-level redundancy and LLT doesn't use IP it's not going to help VCS. IPMP is a replacement / successor to Sun Enterprise Alternate Pathing (and is incompatible with it). In short, AP was for the Enterprise 1 only and provided a means of rerouting the connection to an HBA (whether for storage or network). It is replaced by IPMP and MPXIO (which is better than Veritas DMP but only in Solaris 10+). Bonding isn't really something I expect to see in an Ethernet environment but perhaps that's because I used to do it in an ATM environment years ago. I'll have to look into what modern Ethernet LANs do in this regard. = Dana Hudes UNIX and Imaging group NYC-HRA MIS +1 718 510 8586 Nextel: 172*26*16684 = -Original Message- From: Imri Zvik [mailto:im...@inter.net.il] Sent: Tuesday, May 12, 2009 9:54 PM To: Hudes, Dana; veritas-ha@mailman.eng.auburn.edu Subject: RE: [Veritas-ha] LLT heartbeat redundancy Hi Dana, We were talking about Linux bonding compared to Solaris IPMP. While Solaris link aggregation (apparently) cannot do active/backup, Linux can, and therefore in Linux you can put each link on a separate switch (http://www.kernel.org/pub/linux/kernel/people/marcelo/linux-2.4/Documen tati on/networking/bonding.txt). The whole point of the discussion was how to take the level of redundancy you get for LLT in Linux by using bonding to Solaris by using pure LLT level solution instead of counting on OS specific solutions (As IPMP doesn't actually create a logical interface, LLT cannot use it). -Original Message- From: Hudes, Dana [mailto:hud...@hra.nyc.gov] Sent: Tuesday, May 12, 2009 11:28 PM To: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy Bonding is combining multiple links into one virtual circuit. It isn't usually visible to the OS. IPMP is a bit different. While the result is similar, it's a question of at what level of the tcp/ip stack the combination is done. Bonding is at the media (MAC) layer from an IP perspective (though the OSI model breaks the layers down differently). IPMP is at the network layer. If the combined connections have individual IP addresses, that's multi-path not bonding. There are implications here beyond clustering, including the switch and the router as well as how the increase in bandwidth (if both links are active at once) is achieved. More usual is the use of IPMP to provide standby links: each link has its own IP address and there is an IPMP IP address. Remote hosts use the IPMP address. Often, you will have a primary link which is higher speed / larger bandwidth than the secondary (e.g. gigabit primary and 100 megabit secondary). If you are bonding you must go to the same switch with both links. This provides protection, in the form of degradation of capacity, against port failure on either end of a given path or cable failure of a given path. It does not protect you against switch failure. IPMP, when done properly, has two paths each going to a separate switch. This protects against everything that bonding does while also protecting against switch failure. In the case of switch failure, the standby port becomes active and ARP answers for the virtual IP on that port. The routing protocol, which hopefully is OSPF, quickly redirects the traffic to the new port and you don't even lose your TCP connection due to retransmit (some UDP packets inevitably are lost; an application protocol may have it's own retransmit timer but that's not UDP's bus
Re: [Veritas-ha] LLT heartbeat redundancy
Hi Dana, We were talking about Linux bonding compared to Solaris IPMP. While Solaris link aggregation (apparently) cannot do active/backup, Linux can, and therefore in Linux you can put each link on a separate switch (http://www.kernel.org/pub/linux/kernel/people/marcelo/linux-2.4/Documentati on/networking/bonding.txt). The whole point of the discussion was how to take the level of redundancy you get for LLT in Linux by using bonding to Solaris by using pure LLT level solution instead of counting on OS specific solutions (As IPMP doesn't actually create a logical interface, LLT cannot use it). -Original Message- From: Hudes, Dana [mailto:hud...@hra.nyc.gov] Sent: Tuesday, May 12, 2009 11:28 PM To: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy Bonding is combining multiple links into one virtual circuit. It isn't usually visible to the OS. IPMP is a bit different. While the result is similar, it's a question of at what level of the tcp/ip stack the combination is done. Bonding is at the media (MAC) layer from an IP perspective (though the OSI model breaks the layers down differently). IPMP is at the network layer. If the combined connections have individual IP addresses, that's multi-path not bonding. There are implications here beyond clustering, including the switch and the router as well as how the increase in bandwidth (if both links are active at once) is achieved. More usual is the use of IPMP to provide standby links: each link has its own IP address and there is an IPMP IP address. Remote hosts use the IPMP address. Often, you will have a primary link which is higher speed / larger bandwidth than the secondary (e.g. gigabit primary and 100 megabit secondary). If you are bonding you must go to the same switch with both links. This provides protection, in the form of degradation of capacity, against port failure on either end of a given path or cable failure of a given path. It does not protect you against switch failure. IPMP, when done properly, has two paths each going to a separate switch. This protects against everything that bonding does while also protecting against switch failure. In the case of switch failure, the standby port becomes active and ARP answers for the virtual IP on that port. The routing protocol, which hopefully is OSPF, quickly redirects the traffic to the new port and you don't even lose your TCP connection due to retransmit (some UDP packets inevitably are lost; an application protocol may have it's own retransmit timer but that's not UDP's business). = Dana Hudes UNIX and Imaging group NYC-HRA MIS +1 718 510 8586 Nextel: 172*26*16684 = -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Sunday, May 03, 2009 12:18 PM To: Jim Senicka Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > This is not a limitation, as you had two independent failures. Bonding > would remove the ability to discriminate between a link and a node > failure. I didn't understand this one - With bonding I can maintain full mesh topology - No matter which one of the links fails, if a node still has at least one active link, LLT will still be able to see all the other nodes. This achieves greater HA than without the bonding. > My feeling is in the scenario you describe, VCS is operating properly, > and it is not a limitation. Of course it is operating properly - that's how it was designed to work :) I'm just saying that the cluster could be more redundant if it wasn't designed that way :) > If you have issues with port or cable failures, add a low pri connection > on a third network. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
Bonding is combining multiple links into one virtual circuit. It isn't usually visible to the OS. IPMP is a bit different. While the result is similar, it's a question of at what level of the tcp/ip stack the combination is done. Bonding is at the media (MAC) layer from an IP perspective (though the OSI model breaks the layers down differently). IPMP is at the network layer. If the combined connections have individual IP addresses, that's multi-path not bonding. There are implications here beyond clustering, including the switch and the router as well as how the increase in bandwidth (if both links are active at once) is achieved. More usual is the use of IPMP to provide standby links: each link has its own IP address and there is an IPMP IP address. Remote hosts use the IPMP address. Often, you will have a primary link which is higher speed / larger bandwidth than the secondary (e.g. gigabit primary and 100 megabit secondary). If you are bonding you must go to the same switch with both links. This provides protection, in the form of degradation of capacity, against port failure on either end of a given path or cable failure of a given path. It does not protect you against switch failure. IPMP, when done properly, has two paths each going to a separate switch. This protects against everything that bonding does while also protecting against switch failure. In the case of switch failure, the standby port becomes active and ARP answers for the virtual IP on that port. The routing protocol, which hopefully is OSPF, quickly redirects the traffic to the new port and you don't even lose your TCP connection due to retransmit (some UDP packets inevitably are lost; an application protocol may have it's own retransmit timer but that's not UDP's business). = Dana Hudes UNIX and Imaging group NYC-HRA MIS +1 718 510 8586 Nextel: 172*26*16684 = -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Sunday, May 03, 2009 12:18 PM To: Jim Senicka Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > This is not a limitation, as you had two independent failures. Bonding > would remove the ability to discriminate between a link and a node > failure. I didn't understand this one - With bonding I can maintain full mesh topology - No matter which one of the links fails, if a node still has at least one active link, LLT will still be able to see all the other nodes. This achieves greater HA than without the bonding. > My feeling is in the scenario you describe, VCS is operating properly, > and it is not a limitation. Of course it is operating properly - that's how it was designed to work :) I'm just saying that the cluster could be more redundant if it wasn't designed that way :) > If you have issues with port or cable failures, add a low pri connection > on a third network. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
Yes, the jeopardy detection etc has not changed. We've just added support for this new type of topology in which the various LLT links are interconnected (crosslinks). -Original Message- From: John Cronin [mailto:jsc3...@gmail.com] Sent: Friday, May 08, 2009 6:58 AM To: Sandeep Agarwal (MTV) Cc: Jim Senicka; Imri Zvik; veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy This was certainly news to me. In this full-mesh heartbeat network, do we still go into jeopardy if the network links are lost relatively slowly (e.g. if one link on a node is down for more than 16 seconds by default)? On Sun, May 3, 2009 at 1:07 PM, Sandeep Agarwal (MTV) wrote: From 5.0MP3 onwards we do support cross-links. In your example if you had a cable connecting sw1 and sw2 then the failure that you described would be handled and LLT would still have 1 valid link between node 1 and node 4. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Jim Senicka Sent: Sunday, May 03, 2009 9:23 AM To: Imri Zvik Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy LLT is designed to use "jeopardy" to detect the difference between single link fail and dual link fail in most situations. Having a single mesh may remove this capability. Let me check on this with engineering and see if we have any more up to date recommendations -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Sunday, May 03, 2009 12:18 PM To: Jim Senicka Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > This is not a limitation, as you had two independent failures. Bonding > would remove the ability to discriminate between a link and a node > failure. I didn't understand this one - With bonding I can maintain full mesh topology - No matter which one of the links fails, if a node still has at least one active link, LLT will still be able to see all the other nodes. This achieves greater HA than without the bonding. > My feeling is in the scenario you describe, VCS is operating properly, > and it is not a limitation. Of course it is operating properly - that's how it was designed to work :) I'm just saying that the cluster could be more redundant if it wasn't designed that way :) > If you have issues with port or cable failures, add a low pri connection > on a third network. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
This was certainly news to me. In this full-mesh heartbeat network, do we still go into jeopardy if the network links are lost relatively slowly (e.g. if one link on a node is down for more than 16 seconds by default)? On Sun, May 3, 2009 at 1:07 PM, Sandeep Agarwal (MTV) < sandeep_agarw...@symantec.com> wrote: > From 5.0MP3 onwards we do support cross-links. In your example if you > had a cable connecting sw1 and sw2 then the failure that you described > would be handled and LLT would still have 1 valid link between node 1 > and node 4. > > -Original Message- > From: veritas-ha-boun...@mailman.eng.auburn.edu > [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Jim > Senicka > Sent: Sunday, May 03, 2009 9:23 AM > To: Imri Zvik > Cc: veritas-ha@mailman.eng.auburn.edu > Subject: Re: [Veritas-ha] LLT heartbeat redundancy > > LLT is designed to use "jeopardy" to detect the difference between > single link fail and dual link fail in most situations. Having a single > mesh may remove this capability. > > Let me check on this with engineering and see if we have any more up to > date recommendations > > > -Original Message- > From: veritas-ha-boun...@mailman.eng.auburn.edu > [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri > Zvik > Sent: Sunday, May 03, 2009 12:18 PM > To: Jim Senicka > Cc: veritas-ha@mailman.eng.auburn.edu > Subject: Re: [Veritas-ha] LLT heartbeat redundancy > > On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > > This is not a limitation, as you had two independent failures. Bonding > > > would remove the ability to discriminate between a link and a node > > failure. > > I didn't understand this one - With bonding I can maintain full mesh > topology - No matter which one of the links fails, if a node still has > at least one active link, LLT will still be able to see all the other > nodes. > This achieves greater HA than without the bonding. > > > > My feeling is in the scenario you describe, VCS is operating properly, > > > and it is not a limitation. > > Of course it is operating properly - that's how it was designed to work > :) > I'm just saying that the cluster could be more redundant if it wasn't > designed that way :) > > > If you have issues with port or cable failures, add a low pri > connection > > on a third network. > > > > ___ > Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha > ___ > Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha > ___ > Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu > http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha > ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Wednesday 06 May 2009 19:49:27 Sandeep Agarwal (MTV) wrote: > The lost hb messages are due to problems in Layer 2 connectivity. When I transform the two links into a bonded interface, the lost hb messages goes away (even when I fail over the active NIC back and forth). Is this new feature is supported (I.E. will the tech support support me if I will open a case regarding this feature)? ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
The lost hb messages are due to problems in Layer 2 connectivity. In the cross-linked case as long as 1 link is up on each node LLT should work fine (after a 2s glitch). The llttab file is fine. -Original Message- From: Imri Zvik [mailto:im...@inter.net.il] Sent: Wednesday, May 06, 2009 1:44 AM To: Sandeep Agarwal (MTV) Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Wednesday 06 May 2009 00:55:23 you wrote: > Nothing to add to /etc/llttab > > If you have 50MP3 then it should work. So considering the following llttab file: set-node rac-node1 set-cluster 0 link eth2 eth-00:21:5e:1f:0b:b0 - ether - - link eth3 eth-00:21:5e:1f:0b:b1 - ether - - and that eth2 and eth3 are on the same layer2 subnet (i.e. "sees" each other), I shouldn't get the lost hb messages? ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Wednesday 06 May 2009 00:55:23 you wrote: > Nothing to add to /etc/llttab > > If you have 50MP3 then it should work. So considering the following llttab file: set-node rac-node1 set-cluster 0 link eth2 eth-00:21:5e:1f:0b:b0 - ether - - link eth3 eth-00:21:5e:1f:0b:b1 - ether - - and that eth2 and eth3 are on the same layer2 subnet (i.e. "sees" each other), I shouldn't get the lost hb messages? ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
Nothing to add to /etc/llttab If you have 50MP3 then it should work. -Original Message- From: Imri Zvik [mailto:im...@inter.net.il] Sent: Monday, May 04, 2009 11:12 PM To: Sandeep Agarwal (MTV) Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Monday 04 May 2009 23:51:54 Sandeep Agarwal (MTV) wrote: > Unfortunately we don't have any "more" documentation for this feature. > > Basically, LLT works in the same way except now it can handle the > failure that you described and now we can connect the two switches > that the individual LLT links are configured on. I tried that, but I still get the messages regarding the lost heartbeat in the console. Is there something I need to add in llttab? ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Monday 04 May 2009 23:51:54 Sandeep Agarwal (MTV) wrote: > Unfortunately we don't have any "more" documentation for this feature. > > Basically, LLT works in the same way except now it can handle the > failure that you described and now we can connect the two switches that > the individual LLT links are configured on. I tried that, but I still get the messages regarding the lost heartbeat in the console. Is there something I need to add in llttab? ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
Unfortunately we don't have any "more" documentation for this feature. Basically, LLT works in the same way except now it can handle the failure that you described and now we can connect the two switches that the individual LLT links are configured on. -Original Message- From: Imri Zvik [mailto:im...@inter.net.il] Sent: Sunday, May 03, 2009 11:26 PM To: Sandeep Agarwal (MTV) Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 20:07:29 Sandeep Agarwal (MTV) wrote: > From 5.0MP3 onwards we do support cross-links. In your example if you > had a cable connecting sw1 and sw2 then the failure that you described > would be handled and LLT would still have 1 valid link between node 1 > and node 4. Could you please point me to some more documention regarding this feature? Thanks! ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Sunday 03 May 2009 20:07:29 Sandeep Agarwal (MTV) wrote: > From 5.0MP3 onwards we do support cross-links. In your example if you > had a cable connecting sw1 and sw2 then the failure that you described > would be handled and LLT would still have 1 valid link between node 1 > and node 4. Could you please point me to some more documention regarding this feature? Thanks! ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
Sun Cluster does not support this feature. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Sandeep Agarwal (MTV) Sent: Sunday, May 03, 2009 10:07 AM To: Jim Senicka; Imri Zvik Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy >From 5.0MP3 onwards we do support cross-links. In your example if you had a cable connecting sw1 and sw2 then the failure that you described would be handled and LLT would still have 1 valid link between node 1 and node 4. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Jim Senicka Sent: Sunday, May 03, 2009 9:23 AM To: Imri Zvik Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy LLT is designed to use "jeopardy" to detect the difference between single link fail and dual link fail in most situations. Having a single mesh may remove this capability. Let me check on this with engineering and see if we have any more up to date recommendations -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Sunday, May 03, 2009 12:18 PM To: Jim Senicka Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > This is not a limitation, as you had two independent failures. Bonding > would remove the ability to discriminate between a link and a node > failure. I didn't understand this one - With bonding I can maintain full mesh topology - No matter which one of the links fails, if a node still has at least one active link, LLT will still be able to see all the other nodes. This achieves greater HA than without the bonding. > My feeling is in the scenario you describe, VCS is operating properly, > and it is not a limitation. Of course it is operating properly - that's how it was designed to work :) I'm just saying that the cluster could be more redundant if it wasn't designed that way :) > If you have issues with port or cable failures, add a low pri connection > on a third network. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
>From 5.0MP3 onwards we do support cross-links. In your example if you had a cable connecting sw1 and sw2 then the failure that you described would be handled and LLT would still have 1 valid link between node 1 and node 4. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Jim Senicka Sent: Sunday, May 03, 2009 9:23 AM To: Imri Zvik Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy LLT is designed to use "jeopardy" to detect the difference between single link fail and dual link fail in most situations. Having a single mesh may remove this capability. Let me check on this with engineering and see if we have any more up to date recommendations -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Sunday, May 03, 2009 12:18 PM To: Jim Senicka Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > This is not a limitation, as you had two independent failures. Bonding > would remove the ability to discriminate between a link and a node > failure. I didn't understand this one - With bonding I can maintain full mesh topology - No matter which one of the links fails, if a node still has at least one active link, LLT will still be able to see all the other nodes. This achieves greater HA than without the bonding. > My feeling is in the scenario you describe, VCS is operating properly, > and it is not a limitation. Of course it is operating properly - that's how it was designed to work :) I'm just saying that the cluster could be more redundant if it wasn't designed that way :) > If you have issues with port or cable failures, add a low pri connection > on a third network. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Sunday 03 May 2009 19:22:38 Jim Senicka wrote: > Let me check on this with engineering and see if we have any more up to > date recommendations Thanks! ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
LLT is designed to use "jeopardy" to detect the difference between single link fail and dual link fail in most situations. Having a single mesh may remove this capability. Let me check on this with engineering and see if we have any more up to date recommendations -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Sunday, May 03, 2009 12:18 PM To: Jim Senicka Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > This is not a limitation, as you had two independent failures. Bonding > would remove the ability to discriminate between a link and a node > failure. I didn't understand this one - With bonding I can maintain full mesh topology - No matter which one of the links fails, if a node still has at least one active link, LLT will still be able to see all the other nodes. This achieves greater HA than without the bonding. > My feeling is in the scenario you describe, VCS is operating properly, > and it is not a limitation. Of course it is operating properly - that's how it was designed to work :) I'm just saying that the cluster could be more redundant if it wasn't designed that way :) > If you have issues with port or cable failures, add a low pri connection > on a third network. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Sunday 03 May 2009 19:03:16 Jim Senicka wrote: > This is not a limitation, as you had two independent failures. Bonding > would remove the ability to discriminate between a link and a node > failure. I didn't understand this one - With bonding I can maintain full mesh topology - No matter which one of the links fails, if a node still has at least one active link, LLT will still be able to see all the other nodes. This achieves greater HA than without the bonding. > My feeling is in the scenario you describe, VCS is operating properly, > and it is not a limitation. Of course it is operating properly - that's how it was designed to work :) I'm just saying that the cluster could be more redundant if it wasn't designed that way :) > If you have issues with port or cable failures, add a low pri connection > on a third network. ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
This is not a limitation, as you had two independent failures. Bonding would remove the ability to discriminate between a link and a node failure. My feeling is in the scenario you describe, VCS is operating properly, and it is not a limitation. If you have issues with port or cable failures, add a low pri connection on a third network. -Original Message- From: Imri Zvik [mailto:im...@inter.net.il] Sent: Sunday, May 03, 2009 11:57 AM To: Jim Senicka Cc: veritas-ha@mailman.eng.auburn.edu Subject: Re: [Veritas-ha] LLT heartbeat redundancy On Sunday 03 May 2009 18:25:08 Jim Senicka wrote: > You had 2 failures. No real way to design around that. > GAB "visible" would prevent bad things from occurring. Thank you for the fast response :) Well, In linux I can use the bonding module to aggregate the interfaces and work around this limitation. I've read in this discussion: http://www.mail-archive.com/veritas-ha@mailman.eng.auburn.edu/msg01016.h tml That since 5.0MP3 there is a cross-platform solution (I need this for Solaris 10). Do you happen to know more about this feature? Thanks! P.S. Does anyone knows if Sun Cluster has the same limitation? ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
On Sunday 03 May 2009 18:25:08 Jim Senicka wrote: > You had 2 failures. No real way to design around that. > GAB "visible" would prevent bad things from occurring. Thank you for the fast response :) Well, In linux I can use the bonding module to aggregate the interfaces and work around this limitation. I've read in this discussion: http://www.mail-archive.com/veritas-ha@mailman.eng.auburn.edu/msg01016.html That since 5.0MP3 there is a cross-platform solution (I need this for Solaris 10). Do you happen to know more about this feature? Thanks! P.S. Does anyone knows if Sun Cluster has the same limitation? ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha
Re: [Veritas-ha] LLT heartbeat redundancy
You had 2 failures. No real way to design around that. GAB "visible" would prevent bad things from occurring. -Original Message- From: veritas-ha-boun...@mailman.eng.auburn.edu [mailto:veritas-ha-boun...@mailman.eng.auburn.edu] On Behalf Of Imri Zvik Sent: Sunday, May 03, 2009 11:20 AM To: veritas-ha@mailman.eng.auburn.edu Subject: [Veritas-ha] LLT heartbeat redundancy Hi, As far as I understand the manuals, the LLT heartbeat links should be isolated from each other. Now, consider the following scenario - Node1 is connected with two links, each one to a sperate switch. We will call them li...@node1@sw1 and li...@node1@sw2. Node4 is also connected with 2 links: li...@node4@sw1 and li...@node4@sw2. Now, if Node1 lose li...@node1@sw1 and Node4 lose li...@node4@sw2, they can no longer see each other, although they still have a valid heartbeat link. Am I missing something? Is there a way around this issue? -- imriz ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha ___ Veritas-ha maillist - Veritas-ha@mailman.eng.auburn.edu http://mailman.eng.auburn.edu/mailman/listinfo/veritas-ha