Cluster Hardware Layout
Hello;
We are in the process of setting up a Weblogic cluster with 3 Solaris
servers all running Weblogic 5.1 . Somone has suggested that we isolate
these servers on there own subnet or vlan as the Weblogic servers become
very "chatty". He also recommneded the following:
That each box have a direct connection to EMC shared disk. The EMC shared
disk will be presented as a logical volume to all servers in cluster. (this
is network connection #1)
Each server in cluster should have a Gigabit ethernet interface that is only
for communication between the servers in the cluster (i.e.- multicasting
etc..) (this is network connection #2)
Each server would have a public Gigabit ethernet interface that is listening
for client requests. ( this is network connection #3)
My questions are:
1. Is this configuration possible within the scope of weblogic clustering?
2. Is it practical?
3. If the answers to the above questions are "yes" how can I tell the
different weblogic instances to listen for multicast broadcasts and to
perform memory replications across connection #2 instead of connection #3.
4. EMC has told me that we can carve out some disk space and present it to
multiple machines to be shared disk space, has anyone else done this with
weblogic?
Thanks for your time.
-Bob
"Bob Olson" <[email protected]> wrote:
>Hello;
>
>We are in the process of setting up a Weblogic cluster with 3 Solaris
>servers all running Weblogic 5.1 . Somone has suggested that we isolate
>these servers on there own subnet or vlan as the Weblogic servers become
>very "chatty". He also recommneded the following:
If you are using a switch which has many other hosts and servers connected to it besides the 3 Weblogic Servers
you will want to isolate Multicast traffic by putting the 3 WebLogic servers in a separate VLAN. If you have the
3 servers connected to a shared medium such as bus network topology you will probably want to use VLSM and
routers to create a seperate subnet for the 3 servers (This will isolate Multicast traffic to just the 3 WebLogic
servers) vs. the whole network receiving the Multicast traffic.
>
>That each box have a direct connection to EMC shared disk. The EMC shared
>disk will be presented as a logical volume to all servers in cluster. (this
>is network connection #1)
This sounds like a Storage Area Network (SAN) type solution. Hopefully it will also have transparent
failover and backup.
>
>Each server in cluster should have a Gigabit ethernet interface that is only
>for communication between the servers in the cluster (i.e.- multicasting
>etc..) (this is network connection #2)
This might be overkill for three machines. Most of the initial traffic for Multicast is when the
WebLogic servers start up. After that, there are occational hearbeats that are sent out.
Hot deploying and undeploying objects could increase this traffic but it probably would not come close
to utilizing 1% of the Gigabit ethernet pipe.
>
>Each server would have a public Gigabit ethernet interface that is listening
>for client requests. ( this is network connection #3)
>
From my understanding a single instance of WebLogic can only bind to a single IP address.
I don't think you need 3 physical connections. If you load was great I probably would use
single Gigabit switch with each of the following connected to a port on the switch:
SAN (EMC), WebLogicServer1 ,WebLogicServer2 ,WebLogicServer3, etc.
>My questions are:
>
>1. Is this configuration possible within the scope of weblogic clustering?
>2. Is it practical?
>3. If the answers to the above questions are "yes" how can I tell the
>different weblogic instances to listen for multicast broadcasts and to
>perform memory replications across connection #2 instead of connection #3.
I don't think this is possible currently
>4. EMC has told me that we can carve out some disk space and present it to
>multiple machines to be shared disk space, has anyone else done this with
>weblogic?
Haven't done this but read about it.
>
>Thanks for your time.
You welcome.
>
>-Bob
>
>
Hope This Helps,
Abraham Kang
Similar Messages
-
Switch stacking - recommended hardware layout?
We are a distributed company with a number of sites. I am in the process of redesigning the majority of the comms racks for each site and had a specific question around hardware layout. For info, the majority of edge switches in use are Catalyst 2960S.
My preferred method is to install the switches between the patch panels and use short patch leads (25cm) to connect between. The switches are stacked with appropriate cables at the rear. This obviously makes it much easier to trace cables and I personally feel it looks way better.
Example:
---24 Port Panel---
---Blank---
---48 Port Switch---
---Blank---
---24 Port Panel---
However, others in the IT team say that Cisco's recommended layout is to group all stacked switches together and patch out from there.
I am hoping for a definitive response from a Cisco tech, as well as feedback from others in the group.
ThanksIts all matter of preference. Either solution will work fine. If you have enough space in the rack and the stacking cables will reach from the top switch to the bottom you can leave a blank space between them. As far as I know, there is no recommendation from Cisco in this regard. If you leave space between the switches, its defiantly easier to plug and unplug cables from the ports. As matter of fact I do the same with chassis if the chassis is not full.
HTH -
Oracle 9i Real Application Cluster Hardware Recommandation?
Is there any recommandation for hardware setup for Oracle 9i?
I am looking for a hardware setup to support Oracle 9i Real
Application Cluster.
Thanks in advances
LaurenceIn this weekend I will begin that odessy...
I will give some feedback by monday!
FS -
Need some assistance. I have 2 Ironport in a cluster running. I need to swap hardware one at a time. One device is already disconnected form the cluster. I download the config and uploaded to the new hardware but i always get an error while uploading config to new hardware. Any assistance will be very appreciated.
FYI. new hardware is running the same code as the old hadrware being swapped outKen Stieers wrote:
> I'm betting that there's a guid or key or something that the clustering uses...
Indeed there is, it's the appliance's serial number. You can see it in the config file, where machine-specific settings are in sections identified by each unit's serial number.
Ken's remedy is spot on: remove the old unit from the cluster, give the new one just enough of a config for it to use the network, and join it to the cluster. It will inherit all its other settings from the cluster. If there are any machine-specific settings beyond the normal stuff like host names and IP addresses, then you'll need to add then back manually. I've done this many times, both for replacing failed hardware (fortunately a rare occurrence), and for bringing new units online. Fortunately, all our settings beyond the must-be-local parts are defined at the cluster level.
It's worth pointing out that removing one node from a two-node cluster does not cause the cluster to cease to exist. It's perfectly reasonable for a cluster to have only one member.
++Don -
Hyper-V 2012 R2 Cluster - Drain Roles / Fail Roles Back
Hi all,
In the past when I've needed to apply windows updates to my 3 Hyper-V cluster nodes I used to make a note of which VM's were running on each node, then I'd live migrate them to one of the other cluster nodes before pausing the node I need to work on and
carry out the updates, once I finished installing the updates I'd then simply resume the node and live migrate the VM's back to their original node.
Having recently upgraded my nodes to Windows 2012 R2 I decided to use the new functionality in Failover Cluster Manager where you can pause & drain a node of its roles, perform the updates/maintenance, and then resume & fail roles back to the node,
unfortunately this didn't go as smoothly as I'd hoped, for some reason it seems like the drain/fail back decided to be cumulative rather than one off jobs per-node ... hard to explain, hopefully the following will be clear enough if the formatting survives:
1. Beginning State:
Hyper1 Hyper2 Hyper3
VM01 VM04 VM07
VM02 VM05 VM08
VM03 VM06 VM09
2. Drain Hyper1:
Hyper1 Hyper2 Hyper3
VM04 VM01
VM05 VM02
VM06 VM03
VM07
VM08
VM09
3. Fail Roles Back:
Hyper1 Hyper2 Hyper3
VM01 VM04 VM07
VM02 VM05 VM08
VM03 VM06 VM09
4. Drain Hyper2:
Hyper1 Hyper2 Hyper3
VM01 VM04
VM02 VM05
VM03 VM06
VM07
VM08
VM09
5. Fail Roles Back:
Hyper1 Hyper2 Hyper3
VM01 VM07
VM02 VM08
VM03 VM09
VM04
VM05
VM06
6. Manually Live Migrate VM's back to correct location:
Hyper1 Hyper2 Hyper3
VM01 VM04 VM07
VM02 VM05 VM08
VM03 VM06 VM09
7. Drain Hyper3:
Hyper1 Hyper2 Hyper3
VM01 VM04
VM02 VM05
VM03 VM06
VM07
VM08
VM09
8. Fail Roles Back:
Hyper1 Hyper2 Hyper3
VM01
VM02
VM03
VM04
VM05
VM06
VM07
VM08
VM09
9. Manually Live Migrate VM's back to correct location:
Hyper1 Hyper2 Hyper3
VM01 VM04 VM07
VM02 VM05 VM08
VM03 VM06 VM09
Step 8 was a rather hairy moment, although I was pleased to see my cluster hardware capacity planning rubber stamped, good to know that if I were ever to loose 2 out of 3 nodes everything would keep ticking over!
So, I'm back to the old ways of doing things for now, has anyone else experienced this strange behaviour?
Thanks in advance,
BenHi,
Just want to confirm the current situations.
Please feel free to let us know if you need further assistance.
Regards.
We
are trying to better understand customer views on social support experience, so your participation in this
interview project would be greatly appreciated if you have time.
Thanks for helping make community forums a great place. -
Show the company I work for that we should have one cluster
The company I work for have several projects where we are going to use RAC. The issue I'm having is they want each database to be in its own cluster hardware. I'm trying to show then that we should cluster all the hardware together and make small RAC database in the one cluster.
Does any know of anyone that has written a go doc how why you should have one cluster versus multiple clusters.
Example of what they are thinking of doing note this is all separate hardware cluster1 will not be cluster with cluster2 I'm try to show them if we cluster it all together we get more power and may not need so much hardware:
Project 1
cluster1_dev
2 nodes
cluster2_qa
4 nodes
cluster3_prd
4 nodes
Project 2
cluster2_dev
2 nodes
cluster2_qa
4 nodes
cluster2_prd
4 nodes
What I’m proposing is the following setup:
Cluster_dev
3 nodes
Project1 instance cluster on all 3 nodes
Project2 instance cluster on all 3 nodes
Cluster_qa
6 nodes
Project1 instance cluster on all 6 nodes
Project2 instance cluster on all 6 nodes
Cluster_prd
6 nodes
Project1 instance cluster on all 6 nodes
Project2 instance cluster on all 6 nodesI am always impressed by choosing rac without having expressed what types of failures to be covered.
maybe you really don't need rac ...
moreover, rac is imho really not 2 nodes oriented but well grid oriented.
a lot of small server in the same cluster is for me the way to go or the the way to think.
Services features were introduced to allow one rac db to be shared by multiple applications. -
HT5633 Apple TV Bluetooth Keyboard Layout
Hi,
how can we change the keyboard layout.
It doesn't recognize the AZERTY layout and use QWERTY instead...
CheersOnly English QWERTY is supported.
"Apple TV uses the U.S. English keyboard layout regardless of the keyboard's hardware layout."
I had some friends that ran into this. Probably best to submit feedback at http://www.apple.com/feedback/.
- Cheers -
Client invoke WebService in cluster
Dear All,
I have a problem about call the web service from VB client
I am using Weblogic 7.0 in with 4 managed server in cluster. We will develop a webservice for internal client in VB to use. I have question about how to specific the webservice endpoint address so that the VB client can experience failover and load balancing.
Since we do not have external load balancer and only use IP addresses for connection, how the VB client can connected to the webservice by specifiing the cluster address ? Our cluster address is contains 4 IP address. (e.g. 10.98.103.1:7103,10.98.103.2:7103,10.98.103.3:7103,10.98.103.4:7103)
Thank you!
Best Regards,
Boristo my understanding you do need any kind of loadbalancer in front of WL cluster:
- hardware loadbalancer
- wls pluging for webserver (maybe there is one already?)
- dns round robin
- proxy servlet that comes with WLS (IMHO not supported for production)
there is no magic way for http requests to be distributed in a WLS cluster, that is
why you need a loadbalancer. btw this is different for ejbs, there the loadbalancing
is generated into the client stubs.
so unless you want to program/generate the behaviour in VB you have to thing about a lb.
regards,
Barossa -
This is really annoying. Although my mother language is Portuguese, for which I have a Portuguese-layout Apple Bluetooth keyboard, I have many emails to which I have to reply and write in English. However, when I select the English language in order to use auto-correct options, spelling, etc., the HARDWARE keyboard layout changes into an English keyboard, and it is completely different from what is indicated in the keys. That is: I have to spend a HUGE amount of time trying to find the proper character.
Worse: you cannot tell the English keyboard to have the layout of a Portuguese keyboard, which would be the way to solve this problem. Therefore, I could write in Portuguese AND in English with the same hardware keyboard layout.
And this is only a matter of software, because if you select the English language you have 12 (twelve!) hardware layouts available, NONE Portuguese, which is spoken, and used, by some 240 million people worldwide.
Please do correct this, with the possible urgency.
Thank you.ruibvale wrote:
if you select the English language you have 12 (twelve!) hardware layouts available, NONE Portuguese, which is spoken, and used, by some 240 million people worldwide.
By all means send feedback to Apple. I think part of the problem is that Apple considers the US and US International PC layouts to fit the needs of Portuguese users in Brazil and has just neglected the "Portuguese" layout used in Portugal. There is no reason they could not add it to the list. Hopefully more people will raise the issue with them (yours is the only message I have ever seen about it thus far). -
Hi, Newbie, I am planning for the following X64 Hardware .. is this sane
Hi All,
I am a newbie, and from what i read in Sun Cluster Hardware guide, i have decided to get the following hardware ( with what i can afford :) )
2 AMD 64 Servers with following configuration :
1 AMD 64 X2 Processor
2 GB Ram
3 SATA Hard drives ( internal )
( Nvidia Gigabit Ethernet nge0 )
(Intel Gigabit Ethernet e1000g0 )
(Intel Quad Ethernet , 4 sockets of 100 MBps each )
And for storage :
and one more old AMD 64 machine with following configuration :
it has 3 Ethernet cards and 4 SATA harddrives ( of 160 GB each ).
What i want to do is, as follows :
Disk 1 Disk 2
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
======== ========
Disk 3 Disk 4
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
| 28 Gig | + | 28 Gig | -----> RAID 0 ( 56 Gig )
======== ========
| RAID 0 ( 56 Gig ) | + | RAID 0 ( 56 Gig ) | ----------> RAID 1 56 Gig Mirror ( md1 )
| RAID 0 ( 56 Gig ) | + | RAID 0 ( 56 Gig ) | ----------> RAID 1 56 Gig Mirror ( md2 )
| RAID 0 ( 56 Gig ) | + | RAID 0 ( 56 Gig ) | ----------> RAID 1 56 Gig Mirror ( md3 )
| RAID 0 ( 56 Gig ) | + | RAID 0 ( 56 Gig ) | ----------> RAID 1 56 Gig Mirror ( md4 )
| RAID 0 ( 56 Gig ) | + | RAID 0 ( 56 Gig ) | ----------> RAID 1 56 Gig Mirror ( md5 )
================ ================
And to export all md1,md2,md3,md4,md5 as shared hard drives, through iSCSI.
So both servers , can be iSCSI initiators, and import 5 disk drives. I know performance will be hit , but for learning this should be enough.
NOW Please, tell me if i have any major pitfalls and this setup will not work, because i don't want to bring this and setup this stuff, later to only realize that some thing in this is not supported.
Also, help me on whether, to get two unmanaged Ethernet switch... is this enough .. i am confused here, tell me also how to complete this configuration .
Please help me Cluster guru's , I await for all valuable inputs and guidance on this. also note that i am a sys admin, but have no knowledge of sun clustering, and very less knowledge of clustering in general.
-- Chandan
Message was edited by:
chandan.maddannaHey Jakub, but is VMWare is a workable solution ?
i mean if i am able to do a two node full fledged cluster and try out some data service agents, for example with apache, oracle RAC etc .. that that will do....
btw, do you have any idea as of when iSCSI would be supported by sun clusters, i mean have you seen any road map for this, or anyone saying that it will be supported shortly or something like that.
because i am thinking if this is just a matter of days or couple of months that i can wait, because i can really have a performing cluster, to test out all my fancies..
nice speaking to you, Thanks for all your info on this.. I thought i would hardly get any valid inputs. hope to hear from ya soon..
-- Chandan -
Hardware : learn SCSI RAC @ Home
Hi Everyone,
First: I dont want to use vmware - i want to build a real hardware rac.
Second: I dont want to use Firewire - i want to build a rac with SCSI :o)
my problem is that i have no special cluster hardware knowledge.
i know i need a shared disk (scsi) and i need a second private network link for hearbeats.
the network link is no problem,
but im not sure what hardware for the shared disk i have to use.
my idea is:
PCI SCSI 160 (ie any Adaptec) Card with external Interface in every Computer and to use a external scsi disk as shared storage.
is this correct or are there special requirements on scsi controller and external scsi disk?
keep in mind: im a student with only a little money - therefore it has to be a cheap sollution.
Another topic is the software - i will try RAC on 10g R2 on Win2003 or Suse Linux Enterprise Server 9 - any special hints?
Thanks for any Help
Greetings from Hamburg
AndreasNothing new under the SUN here.
What you will require is two SCSI HBA's that should be a differential SCSI these days everything is practically LVD (Low Voltage Differential). Be sure you don't use a RAID controller because if you do you need to find out if they are so called "cluster" aware. You don't need nor want hardware RAID in this situation, let Oracle take care of that.
You do need to set the scsi-id on the scsi adapters differently from each other. Default they are on scsi ID 7. Then all you need is a simple casing for an external harddrive (or more expensive for external harddrives) and you are off. You can also use some elbow grease and don't bother about the harddisk enclosure and just wire up an extra powersupply for your external drives (externally of course) and connect those to your drive(s) and run an internal standard scsi cable between them. I do agree that that's really dodgy but for an @home solution a great deal of fun.
I run Oracle on two SUN 420R's with two D1000 SCSI enclosures between them and it's just great works really well for home experiment.
But if you want production ready environment than the idea of shared SCSI is not the way to go, in that case let your boss dig in his pockets for a change and let him flick out 20g and you have a very nice SAN based storage solution filled with FC-SAS disks and if he's willing to be more he can have FC-disks. -
Is it possible to upgrade graphics card in my S410?
I've bought a IdeaPad S410, equipeed with i5-4210U 1.70 GHz , Intel HD Graphics (Intergrated) and Radeon R5 M230. I would like to know if it's possible to upgrade my R5 M230 to a better Graphics card, perhaps R7 M260?
Any reply is greatly appreciated, Thanks in advance..hi MizARe3,
Looking at the S410 hardware layout (page 59-60 of the hardware maintenance manual), the CPU and the discrete Radeon R5 M230 are both integrated to the motherboard, thus, you can't separately upgrade the discrete GPU.
- Link to picture
- Link to picture
Regards
Did someone help you today? Press the star on the left to thank them with a Kudo!
If you find a post helpful and it answers your question, please mark it as an "Accepted Solution"! This will help the rest of the Community with similar issues identify the verified solution and benefit from it.
Follow @LenovoForums on Twitter! -
What is the keyboard on N97 Mini 0585087
Hello!
If somebody has an N97 Mini with product code 0585087, 0585116 or 0591946 (they are the EMEA_SWISS variants), can you tell me what keyboard it has? QUERTY? Or QUERTZ?
Thanks in advance,
Gilles.Hi!
OK. Understood. The reason why I ask is that my phone is PC 0591929 and was a Swiss (QWERTZ) keyboard (it is refered on the list as Country Variant Switzerland CH), and the latest firmware available for it is 10.0.20.
For the PC 0585087, the latest firmware is 11.0.45, which is what I would like to upgrade to... but that phone isn't listed as Switzerland CH, rather it is listed as EMEA_SWISS_QW, which might imply QUERTY keyboard rather than QUERTZ.
Since on the N97 Mini, the keyboard is hardware layout, I don't want to upgrade to a firmware which would change the layout of the keyboard compared to what is physically my phone's layout...
Gilles -
Dialog Instance Java is not up after restoring the DB from Stand alone
Dear Gurus,
We are migrating our Stand alone Production (Dual Stack) system to Cluster hardware and we donu2019t have any dialog instances for Standalone Production system.In the process we have taken the backup from the stand alone system and restored in cluster.
After that, the dialog instance Java in is not coming up but central instance java was working fine
As per our analysis, due to only one entry is there for java in the stand alone database and the same database backup is restored in the cluster, the DI instance java is not coming up. Please any body give us the solution to resolve this issue.
Your earliest reply is very much appreciated.
Regards
KaliDear Manoj,
Thanks for your reply.
My Problem was solved.
As per SAP , it won't suggest or support the DB refresh. so did export and import then it was working fine. -
Creating a DLL in LabVIEW containing a function using pointer to pointer
Hi!
I´m facing the problem to create a DLL from LabVIEW containing the following C-function:
functionF(structS** arrayA) ,
where:
structS is a struct of various data-types (e.g. int)
arrayA is an C-array of structS
How can this be done? Any ideas?
Thank you in advance for your supportThis cannot be done in LabVIEW. LV doesn't have the data type "Pointer" as C has.
The only workaround I can think of is to use an array of strings in LV and in the interface define the parameter as "by reference". In LV you have to typecast the wire with the string to a cluster withthe layout of structS.
As far as I know an array of strings is represented as an array of pointer to strings. This is because strings are variable in length and therefore LV uses this method storing the pointers to the strings in the array and not the elements as usual.
You have to read the document about memory and data types of LabVIEW. In elder versions this was in the LabVIEW Bookshelf but I haven't found it in short in the newer version.
Waldemar
Using 7.1.1, 8.5.1, 8.6.1, 2009 on XP and RT
Don't forget to give Kudos to good answers and/or questions
Maybe you are looking for
-
Hi, First post here and my first Mac. I have a Brother 5150D laser printer currently being used on a Windows l/top. A great printer and I would like to link it up to the MB Pro. I have tried a few sites but no joy in getting a driver for linking it t
-
my number is 070-xxxx-xxxx...and I cant log into get apps or anything tell i give a 7 digit number...which I dont have. Whats the work around as I have already added my adress, zip APO AP etc etc ????
-
WCF Data services remove miliseconds from DateTime when expand
I get some strange behavior, when using WCF Data Services 5.6. In my case, I have table, with 1 column set with Concurrency=Fixed, and this column hold date time field from database, updated each time when row is edited. In case I just retrieve entit
-
Hi How to pay single check to multiple vendors
-
BPEL or OSB for reading data from db and then writing to another db
Hi All, I have a requirement of reading the data from a table in one database and then writing into the table in another databas.I dont need any transformation on the data. I understand that i can do this by creating a BPEL process with DB adapters.