Re: Server Group node
To perform the test, I opened 2 user tools, pointing each to one of the 2 servers in the server group, opened the sample classes form and performed a search. I then closed the user tools one at a time. I agree the the timestamps are strange. I checked the time on the servers and they are the same, they are both synced to the same ntp source. Today I see different results (using the same user tool to log into both servers): *Server1* /* Mon Jan 17 2011 10:49:06.5603 */ LOGIN testfloater /* Mon Jan 17 2011 10:49:06.5675 */ FLOAT GRANT WRITEtestfloater (1 of 16500 write) /* Mon Jan 17 2011 10:49:06.6689 */ LOGOUT testfloater /* Mon Jan 17 2011 10:49:37.7920 */ FLOAT RELEASEtestfloater (0 used of 16500 write) *Server2* /* Mon Jan 17 2011 10:47:09.0125 */ LOGIN testfloater /* Mon Jan 17 2011 10:49:15.3799 */ FLOAT GRANT WRITEtestfloater (1 of 16500 write) /* Mon Jan 17 2011 10:49:15.4756 */ LOGOUT testfloater /* Mon Jan 17 2011 10:49:37.8820 */ FLOAT RELEASEtestfloater (0 used of 16500 write) Seperate user tools for each server: *Server1* /* Mon Jan 17 2011 10:53:48.5218 */ LOGIN testfloater /* Mon Jan 17 2011 10:53:48.5272 */ FLOAT GRANT WRITEtestfloater (1 of 16500 write) /* Mon Jan 17 2011 10:53:48.6231 */ LOGOUT testfloater /* Mon Jan 17 2011 10:54:25.6266 */ FLOAT RELEASEtestfloater (0 used of 16500 write) *Server2* /* Mon Jan 17 2011 10:53:57.2278 */ LOGIN testfloater /* Mon Jan 17 2011 10:53:57.2613 */ FLOAT GRANT WRITEtestfloater (1 of 16500 write) /* Mon Jan 17 2011 10:53:57.3680 */ LOGOUT testfloater /* Mon Jan 17 2011 10:54:20.0711 */ FLOAT RELEASEtestfloater (0 used of 16500 write) Strange... Axton Grams The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. On Mon, Jan 17, 2011 at 10:02 AM, Misi Mladoniczky m...@rrr.se wrote: Hi Axton, An interesting thing is that the first server allocates 2 of 16500, and the next server 1 of 16500. To me the timestamps are strange, or the servers are not synchronized. I have seen these chronologically strange events on a number of different RRR|Licnese customers using server groups. The logouts (RELEASE) seems to happen very close to each other. How exactly did you perform this test? And to David Easter: Is it supported and/or recommended to have the same user accessing multiple servers? Best Regards - Misi, RRR AB, http://www.rrr.se Products from RRR Scandinavia (Best R.O.I. Award at WWRUG10): * RRR|License - Not enough Remedy licenses? Save money by optimizing. * RRR|Log - Performance issues or elusive bugs? Analyze your Remedy logs. Find these products, and many free tools and utilities, at http://rrr.se. Axon wrote From a simple test I was able to allocate 2 floating licenses across 2 remedy servers for the same user where the Remedy servers are in the same server group: From Server 1 aruser.log file: /* Fri Jan 14 2011 16:40:51.1490 */ FLOAT GRANT WRITEtestfloater (2 of 16500 write) /* Fri Jan 14 2011 16:41:28.9884 */ LOGIN testfloater From Server 2 aruser.log file: /* Fri Jan 14 2011 16:41:44.0675 */ FLOAT GRANT WRITEtestfloater (1 of 16500 write) /* Fri Jan 14 2011 16:43:28.3593 */ LOGIN testfloater On logout aruser.log shows the following: From Server 1: /* Fri Jan 14 2011 16:46:42.1756 */ FLOAT RELEASEtestfloater (1 used of 16500 write) From server 2: /* Fri Jan 14 2011 16:46:42.2598 */ FLOAT RELEASEtestfloater (0 used of 16500 write) ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Folks, David Easter's message was correct. The system stores floating user information in a temp table and coordinates things across servers. Now, there have been several postings of logs (the one here and a couple of others that seem to show something slightly different) but in none of those cases was there a look at the allocated number of licenses or looking at what users have what type of license. The system keys current operation out of what is in memory as much as possible and doesn't delay your interaction to always check the coordination table -- unless it is at a point of possibly denying access and then it will check first. The idea is that if all seems OK, just let things go but if there seems to be a limit, then take the time to do a check real time. So, you will find that each system you talk to does the Grant Write line. That is that system saying it is at a point where you will be granted a write token. AND YOU ARE. It just happens that it grants the SAME token on the second system as it did on the first. So, you are holding a token to say you are active on each, but it is the same one. The count in the log message may be off slightly because the system doesn't do all the synchronous coordination just to get that number right. It is usually correct and is always close, but it may be off by a couple when there is close activity. BUT, it will not give the same user multiple licenses. AND, it will not block access to a user when there is available licensing. In fact, the slight delay can occasionally allow an overage of a person or two since we wanted to give efficient access and to err on the side of allowing rather than blocking. Now, there is one more scenario that some users have used to indicate that there are multiple tokens -- but there are not. That is a flaw that if the same user gets connected to several systems and then logs out, there are times when the logout does not occur across all systems correctly. Generally, a user is talking with one system and there is no issue, but if there are to multiple, occasionally, all copies don't get cleaned. The user in that case will KEEP a token until the timeout but does not have 2 tokens. If that user logged in again, they would get the SAME token they already have not a new one. So again, 1 token. The issue that causes this problem is fixed in the 7.6.4 release. It is a situation that is infrequent and is a corner case, but it can occur but can NOT give a user 2 licenses. I hope this helps with any confusion in this area. Doug Mueller From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Friday, January 14, 2011 2:49 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** It might depend on the version of ARServer. Things around licenses in server groups changed in the not too recent past; long enough ago that my assumptions may no longer be correct. This may only apply to servers that pre-date the license form and relied on the license file. From a simple test I was able to allocate 2 floating licenses across 2 remedy servers for the same user where the Remedy servers are in the same server group: From Server 1 aruser.log file: /* Fri Jan 14 2011 16:40:51.1490 */ FLOAT GRANT WRITEtestfloater (2 of 16500 write) /* Fri Jan 14 2011 16:41:28.9884 */ LOGIN testfloater From Server 2 aruser.log file: /* Fri Jan 14 2011 16:41:44.0675 */ FLOAT GRANT WRITEtestfloater (1 of 16500 write) /* Fri Jan 14 2011 16:43:28.3593 */ LOGIN testfloater On logout aruser.log shows the following: From Server 1: /* Fri Jan 14 2011 16:46:42.1756 */ FLOAT RELEASEtestfloater (1 used of 16500 write) From server 2: /* Fri Jan 14 2011 16:46:42.2598 */ FLOAT RELEASEtestfloater (0 used of 16500 write) The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. On Fri, Jan 14, 2011 at 4:26 PM, Easter, David david_eas...@bmc.commailto:david_eas...@bmc.com wrote: ** For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. Just to clarify, one user takes one floating token within a server group - even if there are individual requests from the same user to each server in a server group. So no matter how many different servers in a server group that user accesses, they'll get one token. There is logic in the system that stores logged in users in a table that the various servers can coordinate from. Thus, I do not believe that scenario can actually happen. Maybe I'm misunderstanding the scenario
Re: Server Group node
J..I figured yesterday that I would use a perl script to pull the 'Server-Connection-Name' out of the ar.cfg file..what I was hoping for honestly was a keyword like $SERVER-NODE$ or something 'stupid easy' like that.J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 5:27 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Logs, tcpdump, netstat, and many more utilities can be used to determine this. It's a PITA at times, but doable. Beyond that you can test the nodes individually by connecting to them individually to troubleshoot/diagnose problems. On Thu, Jan 13, 2011 at 4:09 PM, LJ LongWing lj.longw...@gmail.com wrote: ** Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that's specific to a certain node of the group (environment related).and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes..so it would be nice to be able to pinpoint which node, both web and app a user is using. J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 2:15 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I'm setting up server groups. I have the need to know which node I'm currently on when connected through the load balanced name. According to the 7.5 docs.all nodes should have the same 'Server-Name', but each node registers with the group with the 'Server-Connect-Name'. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
You can do a set fields from Configuration ARDBC and the query is something like 'Name' = Server-Connect-Name Kind regards Danny From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 14:56 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** J..I figured yesterday that I would use a perl script to pull the 'Server-Connection-Name' out of the ar.cfg file..what I was hoping for honestly was a keyword like $SERVER-NODE$ or something 'stupid easy' like that.J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 5:27 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Logs, tcpdump, netstat, and many more utilities can be used to determine this. It's a PITA at times, but doable. Beyond that you can test the nodes individually by connecting to them individually to troubleshoot/diagnose problems. On Thu, Jan 13, 2011 at 4:09 PM, LJ LongWing lj.longw...@gmail.com wrote: ** Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that's specific to a certain node of the group (environment related).and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes..so it would be nice to be able to pinpoint which node, both web and app a user is using. J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 2:15 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I'm setting up server groups. I have the need to know which node I'm currently on when connected through the load balanced name. According to the 7.5 docs.all nodes should have the same 'Server-Name', but each node registers with the group with the 'Server-Connect-Name'. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Immediately after sending this I remembered that there is a vendor form in Remedy that allows you to grab ar.cfg parameters through a setfield.the name of the form is 'Configure ARDBC'. So, I have written my workflow to do a set from that form on the 'Server-Connect-Name', this tells me the node J From: LJ LongWing [mailto:lj.longw...@gmail.com] Sent: Friday, January 14, 2011 7:56 AM To: 'arslist@ARSLIST.ORG' Subject: RE: Server Group node J..I figured yesterday that I would use a perl script to pull the 'Server-Connection-Name' out of the ar.cfg file..what I was hoping for honestly was a keyword like $SERVER-NODE$ or something 'stupid easy' like that.J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 5:27 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Logs, tcpdump, netstat, and many more utilities can be used to determine this. It's a PITA at times, but doable. Beyond that you can test the nodes individually by connecting to them individually to troubleshoot/diagnose problems. On Thu, Jan 13, 2011 at 4:09 PM, LJ LongWing lj.longw...@gmail.com wrote: ** Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that's specific to a certain node of the group (environment related).and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes..so it would be nice to be able to pinpoint which node, both web and app a user is using. J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 2:15 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I'm setting up server groups. I have the need to know which node I'm currently on when connected through the load balanced name. According to the 7.5 docs.all nodes should have the same 'Server-Name', but each node registers with the group with the 'Server-Connect-Name'. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Danny Kellett Sent: Friday, January 14, 2011 8:01 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** You can do a set fields from Configuration ARDBC and the query is something like 'Name' = Server-Connect-Name Kind regards Danny From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 14:56 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** J..I figured yesterday that I would use a perl script to pull the 'Server-Connection-Name' out of the ar.cfg file..what I was hoping for honestly was a keyword like $SERVER-NODE$ or something 'stupid easy' like that.J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 5:27 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Logs, tcpdump, netstat, and many more utilities can be used to determine this. It's a PITA at times, but doable. Beyond that you can test the nodes individually by connecting to them individually to troubleshoot/diagnose problems. On Thu, Jan 13, 2011 at 4:09 PM, LJ LongWing lj.longw...@gmail.com wrote: ** Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that's specific to a certain node of the group (environment related).and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes..so it would be nice to be able to pinpoint which node, both web and app a user is using. J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 2:15 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I'm setting up server groups. I have the need to know which node I'm currently on when connected through the load balanced name. According to the 7.5 docs.all nodes should have the same 'Server-Name', but each node registers with the group with the 'Server-Connect-Name'. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
I actually think its worth having some very small text on the home page that tells the user which node they are on. I guess the only thing to remember is that if you don't have the sticky bit set (Session infinity as its sometimes known) then you could get directed to others depending on the LB config. Actually thinking about this, I guess you can not have the sticky bit on the LB in front of the AR Servers as this would mean that all your midtier traffic would only go through one server Food for thought anyway Regards Danny From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 15:28 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Danny Kellett Sent: Friday, January 14, 2011 8:01 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** You can do a set fields from Configuration ARDBC and the query is something like 'Name' = Server-Connect-Name Kind regards Danny From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 14:56 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** J..I figured yesterday that I would use a perl script to pull the 'Server-Connection-Name' out of the ar.cfg file..what I was hoping for honestly was a keyword like $SERVER-NODE$ or something 'stupid easy' like that.J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 5:27 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Logs, tcpdump, netstat, and many more utilities can be used to determine this. It's a PITA at times, but doable. Beyond that you can test the nodes individually by connecting to them individually to troubleshoot/diagnose problems. On Thu, Jan 13, 2011 at 4:09 PM, LJ LongWing lj.longw...@gmail.com wrote: ** Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that's specific to a certain node of the group (environment related).and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes..so it would be nice to be able to pinpoint which node, both web and app a user is using. J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 2:15 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I'm setting up server groups. I have the need to know which node I'm currently on when connected through the load balanced name. According to the 7.5 docs.all nodes should have the same 'Server-Name', but each node registers with the group with the 'Server-Connect-Name'. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade-offs. For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. This statement does not apply to the mid-tier layer though. The J2EE container in use has session information stored that is required to handle requests from a user. If the user is re-directed to a mid-tier server where this session information is not available, the user will fail authentication and will be redirected to the login interface. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. On Fri, Jan 14, 2011 at 10:09 AM, LJ LongWing lj.longw...@gmail.com wrote: ** Danny, According to the documentation, you MUST have session affinity turned on….so it only benefits you to have a server group if you have more than one mid tier node….only benefits the user traffic through that mid tier node at least, the other node can still perform admin operations and such…in addition to being available to other clients. *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Danny Kellett *Sent:* Friday, January 14, 2011 8:32 AM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** I actually think its worth having some very small text on the home page that tells the user which node they are on. I guess the only thing to remember is that if you don’t have the sticky bit set (Session infinity as its sometimes known) then you could get directed to others depending on the LB config. Actually thinking about this, I guess you can not have the sticky bit on the LB in front of the AR Servers as this would mean that all your midtier traffic would only go through one server Food for thought anyway Regards Danny *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *LJ LongWing *Sent:* 14 January 2011 15:28 *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** J *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Danny Kellett *Sent:* Friday, January 14, 2011 8:01 AM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** You can do a set fields from “Configuration ARDBC” and the query is something like ‘Name’ = “Server-Connect-Name” Kind regards Danny *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *LJ LongWing *Sent:* 14 January 2011 14:56 *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** J….I figured yesterday that I would use a perl script to pull the ‘Server-Connection-Name’ out of the ar.cfg file….what I was hoping for honestly was a keyword like $SERVER-NODE$ or something ‘stupid easy’ like that…J *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Axton *Sent:* Thursday, January 13, 2011 5:27 PM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** Logs, tcpdump, netstat, and many more utilities can be used to determine this. It's a PITA at times, but doable. Beyond that you can test the nodes individually by connecting to them individually to troubleshoot/diagnose problems. On Thu, Jan 13, 2011 at 4:09 PM, LJ LongWing lj.longw...@gmail.com wrote: ** Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that’s specific to a certain node of the group (environment related)…and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes….so it would be nice to be able to pinpoint which node, both web and app a user is using. J *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Axton *Sent:* Thursday, January 13, 2011 2:15 PM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I’m setting up server groups. I have the need to know which node I’m currently on when connected through the load balanced name. According to the 7.5 docs…all nodes should have the same ‘Server-Name’, but each node
Re: Server Group node
From Page 8 of 84092-Using A Hardware Load Balancer with AR System.pdf The load balancer acts like a NAT device that routes any TCP or UDP traffic. Since the AR System server uses an ONC-RPC implementation that is layered on top of TCP/IP, AR System server traffic can be load balanced. Because of the nature of the client/server interaction within AR System, the sticky option is required. This reduces the balancing that can occur, but still allows for the spreading of the workload from multiple clients. I'm just using the docs J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Friday, January 14, 2011 10:11 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade-offs. For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. This statement does not apply to the mid-tier layer though. The J2EE container in use has session information stored that is required to handle requests from a user. If the user is re-directed to a mid-tier server where this session information is not available, the user will fail authentication and will be redirected to the login interface. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
I guess its not true load balancing then. Just load sharing. If you have a session infinity set at the LB on in front of the ARServers then there is no difference that having each midtier bypassing the LB and going straight to the ARServer instance. E.g. mt1 to AR1, mt2 to AR2 etc I don't think anyone would want to implement something that could take more than one floating license at one time though Axton right? From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 17:52 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** From Page 8 of 84092-Using A Hardware Load Balancer with AR System.pdf The load balancer acts like a NAT device that routes any TCP or UDP traffic. Since the AR System server uses an ONC-RPC implementation that is layered on top of TCP/IP, AR System server traffic can be load balanced. Because of the nature of the client/server interaction within AR System, the sticky option is required. This reduces the balancing that can occur, but still allows for the spreading of the workload from multiple clients. I'm just using the docs J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Friday, January 14, 2011 10:11 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade-offs. For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. This statement does not apply to the mid-tier layer though. The J2EE container in use has session information stored that is required to handle requests from a user. If the user is re-directed to a mid-tier server where this session information is not available, the user will fail authentication and will be redirected to the login interface. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
The difference between straight to and with LB between mid-tier and App is that if one of the nodes fails, traffic will be automatically be transferred from the broken nodes to the active nodes. From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Danny Kellett Sent: Friday, January 14, 2011 11:00 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** I guess its not true load balancing then. Just load sharing. If you have a session infinity set at the LB on in front of the ARServers then there is no difference that having each midtier bypassing the LB and going straight to the ARServer instance. E.g. mt1 to AR1, mt2 to AR2 etc I don't think anyone would want to implement something that could take more than one floating license at one time though Axton right? From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 17:52 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** From Page 8 of 84092-Using A Hardware Load Balancer with AR System.pdf The load balancer acts like a NAT device that routes any TCP or UDP traffic. Since the AR System server uses an ONC-RPC implementation that is layered on top of TCP/IP, AR System server traffic can be load balanced. Because of the nature of the client/server interaction within AR System, the sticky option is required. This reduces the balancing that can occur, but still allows for the spreading of the workload from multiple clients. I'm just using the docs J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Friday, January 14, 2011 10:11 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade-offs. For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. This statement does not apply to the mid-tier layer though. The J2EE container in use has session information stored that is required to handle requests from a user. If the user is re-directed to a mid-tier server where this session information is not available, the user will fail authentication and will be redirected to the login interface. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Yes I agree but that is fault tolerance not load balancing. From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 18:45 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** The difference between straight to and with LB between mid-tier and App is that if one of the nodes fails, traffic will be automatically be transferred from the broken nodes to the active nodes. From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Danny Kellett Sent: Friday, January 14, 2011 11:00 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** I guess its not true load balancing then. Just load sharing. If you have a session infinity set at the LB on in front of the ARServers then there is no difference that having each midtier bypassing the LB and going straight to the ARServer instance. E.g. mt1 to AR1, mt2 to AR2 etc I don't think anyone would want to implement something that could take more than one floating license at one time though Axton right? From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 17:52 To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** From Page 8 of 84092-Using A Hardware Load Balancer with AR System.pdf The load balancer acts like a NAT device that routes any TCP or UDP traffic. Since the AR System server uses an ONC-RPC implementation that is layered on top of TCP/IP, AR System server traffic can be load balanced. Because of the nature of the client/server interaction within AR System, the sticky option is required. This reduces the balancing that can occur, but still allows for the spreading of the workload from multiple clients. I'm just using the docs J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Friday, January 14, 2011 10:11 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade-offs. For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. This statement does not apply to the mid-tier layer though. The J2EE container in use has session information stored that is required to handle requests from a user. If the user is re-directed to a mid-tier server where this session information is not available, the user will fail authentication and will be redirected to the login interface. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Only 2 groups of people that I can think of that won't care about the floating license multiple allocations: - people that don't use floating licenses - people that don't use (or pay for) floating licenses Chances are that applies to a very small subset of users, if any. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. On Fri, Jan 14, 2011 at 12:00 PM, Danny Kellett dkell...@javasystemsolutions.com wrote: ** I guess its not true load balancing then. Just load sharing. If you have a session infinity set at the LB on in front of the ARServers then there is no difference that having each midtier bypassing the LB and going straight to the ARServer instance. E.g. mt1 to AR1, mt2 to AR2 etc I don’t think anyone would want to implement something that could take more than one floating license at one time though Axton right? *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *LJ LongWing *Sent:* 14 January 2011 17:52 *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** From Page 8 of 84092-Using A Hardware Load Balancer with AR System.pdf The load balancer acts like a NAT device that routes any TCP or UDP traffic. Since the AR System server uses an ONC-RPC implementation that is layered on top of TCP/IP, AR System server traffic can be load balanced. Because of the nature of the client/server interaction within AR System, the sticky option is required. This reduces the balancing that can occur, but still allows for the spreading of the workload from multiple clients. I’m just using the docs J *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Axton *Sent:* Friday, January 14, 2011 10:11 AM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade-offs. For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. This statement does not apply to the mid-tier layer though. The J2EE container in use has session information stored that is required to handle requests from a user. If the user is re-directed to a mid-tier server where this session information is not available, the user will fail authentication and will be redirected to the login interface. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Ø For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. Just to clarify, one user takes one floating token within a server group - even if there are individual requests from the same user to each server in a server group. So no matter how many different servers in a server group that user accesses, they'll get one token. There is logic in the system that stores logged in users in a table that the various servers can coordinate from. Thus, I do not believe that scenario can actually happen. Maybe I'm misunderstanding the scenario... -David J. Easter Manager of Product Management, Remedy Platform BMC Software, Inc. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Friday, January 14, 2011 11:04 AM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Only 2 groups of people that I can think of that won't care about the floating license multiple allocations: - people that don't use floating licenses - people that don't use (or pay for) floating licenses Chances are that applies to a very small subset of users, if any. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. On Fri, Jan 14, 2011 at 12:00 PM, Danny Kellett dkell...@javasystemsolutions.commailto:dkell...@javasystemsolutions.com wrote: ** I guess its not true load balancing then. Just load sharing. If you have a session infinity set at the LB on in front of the ARServers then there is no difference that having each midtier bypassing the LB and going straight to the ARServer instance. E.g. mt1 to AR1, mt2 to AR2 etc I don't think anyone would want to implement something that could take more than one floating license at one time though Axton right? From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORGmailto:arslist@ARSLIST.ORG] On Behalf Of LJ LongWing Sent: 14 January 2011 17:52 To: arslist@ARSLIST.ORGmailto:arslist@ARSLIST.ORG Subject: Re: Server Group node ** From Page 8 of 84092-Using A Hardware Load Balancer with AR System.pdf The load balancer acts like a NAT device that routes any TCP or UDP traffic. Since the AR System server uses an ONC-RPC implementation that is layered on top of TCP/IP, AR System server traffic can be load balanced. Because of the nature of the client/server interaction within AR System, the sticky option is required. This reduces the balancing that can occur, but still allows for the spreading of the workload from multiple clients. I'm just using the docs :) From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORGmailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Friday, January 14, 2011 10:11 AM To: arslist@ARSLIST.ORGmailto:arslist@ARSLIST.ORG Subject: Re: Server Group node ** Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade-offs. For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. This statement does not apply to the mid-tier layer though. The J2EE container in use has session information stored that is required to handle requests from a user. If the user is re-directed to a mid-tier server where this session information is not available, the user will fail authentication and will be redirected to the login interface. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. _attend WWRUG11 www.wwrug.comhttp://www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.comhttp://www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
It might depend on the version of ARServer. Things around licenses in server groups changed in the not too recent past; long enough ago that my assumptions may no longer be correct. This may only apply to servers that pre-date the license form and relied on the license file. From a simple test I was able to allocate 2 floating licenses across 2 remedy servers for the same user where the Remedy servers are in the same server group: From Server 1 aruser.log file: /* Fri Jan 14 2011 16:40:51.1490 */ FLOAT GRANT WRITEtestfloater (2 of 16500 write) /* Fri Jan 14 2011 16:41:28.9884 */ LOGIN testfloater From Server 2 aruser.log file: /* Fri Jan 14 2011 16:41:44.0675 */ FLOAT GRANT WRITEtestfloater (1 of 16500 write) /* Fri Jan 14 2011 16:43:28.3593 */ LOGIN testfloater On logout aruser.log shows the following: From Server 1: /* Fri Jan 14 2011 16:46:42.1756 */ FLOAT RELEASEtestfloater (1 used of 16500 write) From server 2: /* Fri Jan 14 2011 16:46:42.2598 */ FLOAT RELEASEtestfloater (0 used of 16500 write) The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. On Fri, Jan 14, 2011 at 4:26 PM, Easter, David david_eas...@bmc.com wrote: ** Ø “For example, if users have a floating license, they can potentially allocated a floating license on each server that their request is sent to. ” Just to clarify, one user takes one floating token within a server group – even if there are individual requests from the same user to each server in a server group. So no matter how many different servers in a server group that user accesses, they’ll get one token. There is logic in the system that stores logged in users in a table that the various servers can coordinate from. Thus, I do not believe that scenario can actually happen. Maybe I’m misunderstanding the scenario… -David J. Easter Manager of Product Management, Remedy Platform BMC Software, Inc. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Axton *Sent:* Friday, January 14, 2011 11:04 AM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** Only 2 groups of people that I can think of that won't care about the floating license multiple allocations: - people that don't use floating licenses - people that don't use (or pay for) floating licenses Chances are that applies to a very small subset of users, if any. The opinions, statements, and/or suggested courses of action expressed in this E-mail do not necessarily reflect those of BMC Software, Inc. My voluntary participation in this forum is not intended to convey a role as a spokesperson, liaison or public relations representative for BMC Software, Inc. On Fri, Jan 14, 2011 at 12:00 PM, Danny Kellett dkell...@javasystemsolutions.com wrote: ** I guess its not true load balancing then. Just load sharing. If you have a session infinity set at the LB on in front of the ARServers then there is no difference that having each midtier bypassing the LB and going straight to the ARServer instance. E.g. mt1 to AR1, mt2 to AR2 etc I don’t think anyone would want to implement something that could take more than one floating license at one time though Axton right? *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *LJ LongWing *Sent:* 14 January 2011 17:52 *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** From Page 8 of 84092-Using A Hardware Load Balancer with AR System.pdf The load balancer acts like a NAT device that routes any TCP or UDP traffic. Since the AR System server uses an ONC-RPC implementation that is layered on top of TCP/IP, AR System server traffic can be load balanced. Because of the nature of the client/server interaction within AR System, the sticky option is required. This reduces the balancing that can occur, but still allows for the spreading of the workload from multiple clients. I’m just using the docs J *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Axton *Sent:* Friday, January 14, 2011 10:11 AM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** Authentication information is included in each packet to the arserver, so it can run without session persistence. However, there are trade
Re: Server Group node
Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I’m setting up server groups. I have the need to know which node I’m currently on when connected through the load balanced name. According to the 7.5 docs…all nodes should have the same ‘Server-Name’, but each node registers with the group with the ‘Server-Connect-Name’. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that's specific to a certain node of the group (environment related).and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes..so it would be nice to be able to pinpoint which node, both web and app a user is using. J From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of Axton Sent: Thursday, January 13, 2011 2:15 PM To: arslist@ARSLIST.ORG Subject: Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I'm setting up server groups. I have the need to know which node I'm currently on when connected through the load balanced name. According to the 7.5 docs.all nodes should have the same 'Server-Name', but each node registers with the group with the 'Server-Connect-Name'. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
Re: Server Group node
Logs, tcpdump, netstat, and many more utilities can be used to determine this. It's a PITA at times, but doable. Beyond that you can test the nodes individually by connecting to them individually to troubleshoot/diagnose problems. On Thu, Jan 13, 2011 at 4:09 PM, LJ LongWing lj.longw...@gmail.com wrote: ** Well, as we are setting up and configuring the nodes, there is always the possibility that a user may be having trouble, trouble that’s specific to a certain node of the group (environment related)…and without the ability for us to know what app node a user is on, it makes it difficult to know which of the app nodes is causing the problem. In production we have the following architecture Load Balancer 4 Mid-Tier Load Balancer 2 App So, any one of the 4 web nodes could be pointing to any one of the app nodes….so it would be nice to be able to pinpoint which node, both web and app a user is using. J *From:* Action Request System discussion list(ARSList) [mailto: arslist@ARSLIST.ORG] *On Behalf Of *Axton *Sent:* Thursday, January 13, 2011 2:15 PM *To:* arslist@ARSLIST.ORG *Subject:* Re: Server Group node ** Want to share more around what you are trying to accomplish? Why do you need to know the name? On Thu, Jan 13, 2011 at 2:32 PM, LJ LongWing lj.longw...@gmail.com wrote: ** I’m setting up server groups. I have the need to know which node I’m currently on when connected through the load balanced name. According to the 7.5 docs…all nodes should have the same ‘Server-Name’, but each node registers with the group with the ‘Server-Connect-Name’. it is the connect name I want to be able to access through workflow. Any/all suggestions are appreciated. _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ _attend WWRUG11 www.wwrug.com ARSlist: Where the Answers Are_ ___ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug11 www.wwrug.com ARSList: Where the Answers Are
