Two servers, one RAID

RAID newb here. I've searched this forum and found some hints to the answer, but I thought I'd throw it out there anyway, just to be sure.
I have a Promise RAID, only the top enclosure is full. I believe only one controller accesses that top enclosure.
I have two servers. I know only one can access the controller at a time, but if I have two partitions, can each one access its own partition without affecting the other? If so, how do I go about setting that up?
Added bonus: One server is Leopard, the other is Snow Leopard.
It's not a total dealbreaker if this can't be done, I just want to make sure it can't be done before I plan my next step.
Thanks!
Michelle
Message was edited by: ModoReese

There are only two ports in the controller - if you're okay with only accessing the Promise from one connection each you can plug one server into one port and the other into the other.
NOTE: You CANNOT access the same data from two servers at one time. You can't even have the same data mounted on both server simultaneously without risking data corruption.
What you want to look into is LUN masking. You can use LUN masking to determine which unit (based on port) can see what data at any given time. The concept and set-up of this actually relatively easy on the Promise box look through the documentation and it should answer your questions.
What I can't answer for you is whether your current partition scheme is compatible with the LUN masking set-up. I've only set-up Promise boxes from scratch for LUN masking never retroactively against a standing set-up of existing partitions. Sorry for the partial answer but hopefully someone who's done what you're trying to do can address it directly. Worst case scenario involves backing the data off the partitions, reformatting the Promise into LUN maskable volumes and putting the data back onto it.
Good luck,
=Tod

Similar Messages

Two domains + two servers + one static IP address = DNS confusion

I'll try to keep this simple:
I have two domains, two mac mini SL servers, one airport extreme, and multiple static IPs, but only one of them pointing to the router. I've configured DNS successfully for the primary domain (example1.com) and thought that I had set up a second primary zone for the second domain (example2.com) but all external requests point to example1.com.
So the two setups i've tried are:
Reverse zone
1.0.10.in-addr.arpa.
- 10.0.1.200 - example1.com
- 10.0.1.201 - example2.com
Primary zone
example1.com
- example1.com - 10.0.1.200
- example2.com - 10.0.1.201
OR
Primary zone 1
example1.com
- example1.com - 10.0.1.200
Primary zone 2
example2.com
- example2.com - 10.0.1.201
Does anyone know of a good tutorial for hosting multiple servers in an internal network, pointing to one name server internally? I've looked everywhere and cannot make sense of the issue.
Thanks in advance.

After sleeping on it, I came up with a much easier solution that works better for what I'd planned in the first place. I hung a spare router off my gateway and set up a completely different network for the second server and everything is great now. I knew there was a reason I'd signed up for five IPs...
In any event, I don't think I would've been able to accomplish what I was trying to do while using the Airport Extreme as my firewall. I wanted to have separate static IPs for each of the boxes and obviously I couldn't do that with the airport, although it took a while for me to recognize that.
thanks for the suggestion, though. I think I'm going to use that for some subdomains.

Two servers- one ip address

i just got another server to add to my exsisting server running mac os x 10.4
what i want to know is how to tweak both servers so they can share the one public ip address i have. currently a router is keeping the orginal server as a DMZ host. how can i get the ip to share?!?

What kind of router do you have (make/model)?
Is it a simple NAT router with a DMZ "passthrough" to a specific LAN IP or what?
If this is your setup you can "divide" services between two or more servers by using portforwarding in the router, so a couple is served by one machine and others (different services) by another machine.
Beware a DMZ like this is forwarding "all other ports" to the LAN IP machine so you might want to enable a firewall in that machine only allowing the public services to be served (to Internet and LAN depending on setup).

Storage redundancy using two OSX servers - Promise RAID

Hi All,
Currently we run 1 Mac Mini and Server on Mavericks connected via Thunderbolt to a Promise RAID 5 device, it works great. I want some redundancy and have purchased an additional Mac Mini to backup Open directory and DNS as our main server does pretty much everything.
Incase the main server fails what is the best way for the Promise RAID to stay up? I've read you can create a "Thunderbolt Bridge" by connecting both Mac Mini's via thunderbolt cable and one to the RAID??? Is this correct? Or.... purchase a Thunderbolt to Gigiabyte adapter and run the RAID off the network. Any idea how to configure this?
Thanks.
James

Sadly Apple do not provide much if any facilities for building a resilient setup along the lines you are thinking.
The nearest you can get happens to be for Open Directory and DNS. You would setup two servers, one as the DNS master and one as a DNS secondary, and you would setup one server as the Open Directory Master, and the other as the Open Directory Replica. Neither of these require the two servers to share any storage or files, the two servers would send the information between themselves over the network. This only works for those two services.
You couldn't for example have both connected to the promise and act as File Servers. (Both ordinary files sharing or network directories.)
To get nearer to having two servers sharing common storage so that if one goes down the other takes over would require a SAN and not direct attached storage like the Promise. However Apple's software does not support this for file sharing services.

Combine two servers to one

Hello,
We have the system 11.5.7:
AppsTier on one server and the DBTier, concurrent managers and the admin on another server. There is no the load balance for the AppsTier. The load balance requests more than one server. So we are on the Single Node architecture and Not the Multinode architecture, though there are two servers (one for AppsTier and one for DBTier and etc) in the system.
The system has been in production about 3 years. There has never been over loaded due to the user sessions in the system. I believe we should combine the system 11.5.7 on the two servers to one server. We have the server can handle the load with significant free margin in terms of the server resources, and the performance will be improved by eliminating the latencies between the servers in the current system. Another major benefit is that the simplification for the maintenance and the upgrade.
I need to convince the management for consolidating the two servers to one server as described. Do you agree the proposed one server system is better than two servers for our system?
Thanks In Advance.
Sean

i know it will easier to manager one node rathter than multi-node but my point of view is that you will have a complete server to server your (froms and web) and another complete server for your database and concurrent manager.
maybe you and other users have different resones and i am intersted to know is it just becase the management of one server will be easier or are there other resons.
one qustions please is your single server more prowerfull than the other 2 servers or not?
fadi

2 servers and one raid system

HI All,
I've connected two servers, each a different controller, to one LUN of a raid system. All the raid disks (except the spare) are put together inside 1 LUN.
Now I think I remember that this is not a good idea. Can someone explain to me why this is bad (or not) ?
Thanks in advance
Luca

Look for clustering on the forum or rather scsi-initiator-id.

One client and two servers

hi,
how a client can communicate to two servers by socket connection (parallel) ?
here is my code, when i run one server at a time it works but if i run both server programs and then the client i dont get any exception and it does not get my data from servers it generates "press any key to continue..."
server 1 --------------
public static void main(String[] args)
ServerOne so = new ServerOne(7500);
so.accept();
public void accept(){
try{
System.out.println("Server");
Socket s;
while(true)
s = sock.accept();
}catch()...
server 2 -----
exact like server 1
klient---------------------------
public static void main(String[] args)
try{
System.out.println("Client");
MyThread MT = new MyThread();
MT.startThread();
Socket s = new Socket("127.0.0.1",7500);
InputStream is = s.getInputStream();
BufferedReader in = new BufferedReader(new InputStreamReader(is));
while(in.ready())
String str = in.readLine();
System.out.println(str);
s.close();
}catch(IOException e){e.printStackTrace();}
mythread --------------------
public void run()
? try
Socket s = new Socket("127.0.0.1",1234);
InputStream is = s.getInputStream();
BufferedReader in= new BufferedReader(new InputStreamReader(is));
while(in.ready())
String str = in.readLine();
System.out.println(str);
s.close();
catch(Exception e){System.out.println(e.toString());}
Message was edited by:
cyrus

You cannot run two servers on exactly the same socket. Are you aware of this?
At least the port number has to be different.
// Server 1:
try {
    serverSocket = new ServerSocket(4444);
} catch (IOException e) {
    System.out.println("Could not listen on port: 4444");
    System.exit(-1);
Socket clientSocket = null;
try {
    clientSocket = serverSocket.accept();
} catch (IOException e) {
    System.out.println("Accept failed: 4444");
    System.exit(-1);
// now the connection is established
// do the same for the other server, but for another port
// The client has to explicitly call each Server, to establish each connection.

How to set two radius servers one is window NPS another is cisco radius server

how to set two radius servers one is window NPS another is cisco radius server
when i try the following command, once window priority is first , i type cisco radius user name, it authenticated fail
i can not use both at the same time
radius-server host 192.168.1.3 is window NPS
radius-server host 192.168.1.1 is cisco radius
http://blog.skufel.net/2012/06/how-to-integrating-cisco-devices-access-with-microsoft-npsradius/
conf t
no aaa authentication login default line
no aaa authentication login local group radius
no aaa authorization exec default group radius if-authenticated
no aaa authorization network default group radius
no aaa accounting connection default start-stop group radius
aaa new-model
aaa group server radius IAS
server 192.168.1.1 auth-port 1812 acct-port 1813
server 192.168.1.3 auth-port 1812 acct-port 1813
aaa authentication login userAuthentication local group IAS
aaa authorization exec userAuthorization local group IAS if-authenticated
aaa authorization network userAuthorization local group IAS
aaa accounting exec default start-stop group IAS
aaa accounting system default start-stop group IAS
aaa session-id common
radius-server host 192.168.1.1 auth-port 1812 acct-port 1813
radius-server host 192.168.1.2 auth-port 1812 acct-port 1813
radius-server host 192.168.1.3 auth-port 1645 acct-port 1646
radius-server host 192.168.1.3 auth-port 1812 acct-port 1813
privilege exec level 1 show config
ip radius source-interface Gi0/1
line vty 0 4
authorization exec userAuthorization
login authentication userAuthentication
transport input telnet
line vty 5 15
authorization exec userAuthorization
login authentication userAuthentication
transport input telnet
end
conf t
aaa group server radius IAS
server 192.168.1.3 auth-port 1812 acct-port 1813
server 192.168.1.1 auth-port 1812 acct-port 1813
end

The first AAA server listed in your config will always be used unless/until it becomes unavailable. At that point the NAD would move down to the next AAA server defined on the list and use that one until it becomes unavailable and then move to third one, and so on.
If you want to use two AAA servers at the same time then you will need to put a load balancer in front of them. Then the virtual IP (vip) will be listed in the NADs vs the individual AAA servers' IPs.
I hope this helps!
Thank you for rating helpful posts!

RAID NAS and Three Authorized Computers Two MAC ONE PC

Hi All,
So I have some time off from work for the holiday. My goal is two have one main library of music that both my kids and I can share.
They have Macbooks and I have a PC. We are all networked here and I have a Dell Raid NAS that I am planning on storing all the music on.
Right now the main library is stored on my Windoze PC and a Briefcase copy is kept on the RAID server. That is for all intents and purposes the main file.
I plan on redirecting Itunes on my PC to look at that library as the main store.
I also plan on having both the kids Mac's look to that store as well.
I do have some questions that I think I have the answer to at least one.
1) if Child A makes a purchase on her machine when it is not physically connected to the RAID server, I believe that Itunes on the macbook will then save it LOCALLY then when I CONSOLIDATE it will move it to the main RAID STORE.
This one I am not sure about......
2) if child A makes a purchase on her machine WHILE connected to the RAID server main store Itunes will store the file on the main RAID STORE. However, when child B's Macbook starts and she starts itunes will her itunes see the file that child A purchased in her list of music.
I guess in the nutshell the question is if multiple CPU's are looking at a MAIN storage library and if one machine makes an addition, to the library. When the other Itunes start will it scan the libray and be updated with the addition?
Thanks for the help
Nick

Let's say, I would like to NOT use the old PPC Mac Mini as the input method for new music.
I would like to leave the iTunes library where it is (on an external HDD) and simply point the new iMac to this library and update it as necessary. I want to know how can it be done by leaving the external HDD connected to the old Mac Mini.
Does it make sense?
Thanks!

Question--Two servers, two ethernet cables, one airport input.

I have two servers, Earthlink and Sbc, How do I hook both of these to the one input of the airport? Arne
powermac g5 Mac OS X (10.4.5)

See the answer already posted.
Don

One Report, Two Servers, Debugging

I have a report that uses multi-select parameters.
It is deployed on two SSRS 2008 R2 Report Servers.
It will run on one but not other.
Are there any kinds of SSRS 2008 R2 debugging tools to find out what is the difference between the two servers?
It may be an ODBC driver but I am not sure.
Thank you.

First we need to identify what is generating the error on the one server. This error occurs when you pass too much for a given data type. For instance, on one server an int parameter value that is generated from a SQL query might result in a bigint value
that overflows the int data type. Or a string passed to a char(10) might exceed 10 characters while it does not on the other server.
What is the query text where these 2 parameters are used?
If you remove these parameters from the data set filter, will the report run without the error?
If so, what are the values for these 2 parameters? (add 2 text boxes to the report to display the parameter values using Join(Parameters!Parameter1.Value,"; "))
Then we need to determine why it is not happening on the other.
Why is it deployed to 2 servers? Is it connecting to different data sources?
What is the difference between the two servers where it is deployed.
"You will find a fortune, though it will not be the one you seek." -
Blind Seer, O Brother Where Art Thou
Please Mark posts as answers or helpful so that others may find the fortune they seek.

Import settings, then I get two servers on 8.6 BE, Deleted one, now I can't log in

I had a base install of 8.6.1 BE. I got the licenses installed and teamed the nics.
I went to the live 8.5.1 stand alone live CUCM server and went to Bulk Admin --> Import/Export. I exported everything available.
I imported that tar file tot eh 8.6.1 BE CUCM server. Then I noticed that there were now two servers listed under System-->Server. One was the name of the server during my build, the other was the name of the imported server and that was listed as a subscriber.
So, I deleted the subscriber. Now after a reboot I can't log in. I get "Database Communication Error."
Please help!

Nevermind!
I have to change the hostname via the command line and reboot. Now everything is working.

How to split mailservices among two servers (no xsan yet)

How can I split the load on our mailserver (getting rid of the daily VM growth and potential crashes / down time) by using TWO servers as opposed to one?
Apparently the root of the cause is too many imapd connections and DirectoryServices then gets in to problems stalling the whole server (EG see my latest 3-4 posts)
SMTP out of the house should not pose a problem, and access for the end users to the (manually moved) imap data location should be handled easily after pointing the users to the second mailserver in WGM.
BUT incoming mail makes me unsure - port 25 can only be pointed to one server on the LAN (naturally , so how do I configure the ("primary") mailserver to make sure that the local mail-delivery (for the users that use the second mailserver) will actually be sent to this second server.
I would think that this should be easily possible - somehow - and that the LMTP service will be the part responsible for this change.
PS.: I still cannot understand why a dual 2 GHz intel server is unable to perform satisfactory - with 10.5.6 server - serving the small amount of IMAP users we have, ca. 250 users.
Currently ps ax | grep imapd | wc -l tells me that I have up to ca. 150 connections when the VM growth warnings starts to show up in the system.log

The server is still pretty much having the default setup, though I have implemented Pterobytes "Frontline Spam Defense", but:
A. I have incresed the number of processes as per an Apple article for "heavily loaded servers":
cat /etc/launchd.conf
limit maxproc 1000 2000
B. I have changed /ets/cyrus.conf
IMAP increased to 5
LMTPUNIX increased to 2
<pre> imap cmd="imapd" listen="imap" prefork=5
imaps cmd="imapd -s" listen="imaps" prefork=0
lmtpunix cmd="lmtpd" listen="/Volumes/Bric-RAID-03-up/Mail/imap/socket/lmtp" prefork=2</pre>
I got this advise from Pterobyte a year ago when I had problems with the performance of the old mailserver (the hw was a G4, the hint helped a lot!)
The timeout you refer to is not set in the config file (also not in a fresh in stall of 10.5.6), it's only listed (as 30) in /etc/imapd.conf.default.
But I presume that if the parameter is not set in my cinf file, then the default files settings are in effect(?)
I was thinking of changing this value just yesterday, but after reading the man imapd.conf I thought that it would not make any diference, since 30 is "the minimum":
timeout: 30
The length of the IMAP server's inactivity autologout timer, in minutes. The
minimum value is 30, the default.
I guess I have to make this chang anyway
The server is a dual 2 GHz intel with 5 GB ram (and still 4 vacant slots).
Today's top output looks like this (the server was rebooted 24 hours ago!)
Mail spool (currently ca. 120 GB) is on an RAID-5 xraid
<pre>Processes: 171 total, 2 running, 169 sleeping... 741 threads 10:57:37
Load Avg: 0.21, 0.16, 0.06 CPU usage: 3.39% user, 3.39% sys, 93.23% idle
SharedLibs: num = 12, resident = 44M code, 1532K data, 3200K linkedit.
MemRegions: num = 16009, resident = 625M + 27M private, 231M shared.
PhysMem: 708M wired, 942M active, 174M inactive, 1826M used, 3294M free.
VM: 18G + 373M 645958(0) pageins, 75(0) pageouts</pre>
I have been running Apples top-guide.pl script for a couple of days to monitor the servers "health" in general as well as I am logging the number of imapd connections every minute.
I dont find the numbers alarming.
During the max load yesterday (between 12.10-12.48) there was a peak of 210 imapd connection (users coming back from lunch and checking mail?).
<pre>Fri Feb 6 12:42:00 CET 2009: 210
Fri Feb 6 12:43:00 CET 2009: 210
Fri Feb 6 12:44:00 CET 2009: 210
Fri Feb 6 12:45:00 CET 2009: 209
Fri Feb 6 12:46:00 CET 2009: 210
Fri Feb 6 12:47:00 CET 2009: 209
Fri Feb 6 12:48:00 CET 2009: 210
</pre>
During these 8 minutes is also when I was logging the highest number of VM growth:
<pre>Feb 6 12:42:35 einstein DirectoryService[26]: Potential VM growth in DirectoryService since client PID: 0, has 875 open references when the warning limit is 500.
</pre>
top-guide.pl log's these vaulues:
<pre>
CPU usage: 300-s Reads/sec: Writes/sec: Net in/sec: Net out/sec: kernel
Time: user sys Idle avg. number MB number MB pkts Mb pkts Mb CPU%
12:46:29 8.0 3.8 88.3 87.4 0 0.0 65 3.1 2138 21.3 2081 11.4 3.7
12:46:39 5.8 3.7 90.6 87.5 4 0.1 30 2.8 2166 21.2 2108 11.2 3.9
12:46:49 5.4 4.5 90.1 87.5 0 0.0 22 3.1 2210 21.4 2180 11.7 4.0
12:46:59 10.3 4.6 85.1 87.3 0 0.0 54 3.1 2123 21.2 2091 11.4 3.8
12:47:09 13.8 5.2 81.0 86.9 0 0.0 46 3.2 2110 21.2 2071 11.4 3.9
12:47:19 4.7 3.1 92.2 87.0 4 0.1 24 2.6 2109 21.1 2073 11.4 3.7
12:47:29 5.0 3.8 91.2 87.6 0 0.0 58 3.3 2136 21.2 2110 11.3 3.8
12:47:39 5.4 3.6 91.0 87.6 35 0.5 61 2.9 2091 21.3 2042 11.4 3.8
12:47:49 7.9 5.9 86.2 87.5 61 0.7 54 4.8 2424 23.8 2409 13.8 4.5
12:47:59 5.2 4.1 90.8 87.4 0 0.0 52 3.7 2224 21.4 2222 12.2 4.0
12:48:09 15.9 5.9 78.2 87.0 0 0.0 25 3.0 2146 21.1 2127 11.8 4.0
12:48:19 7.4 3.4 89.2 87.1 0 0.0 15 2.6 2129 21.2 2111 11.8 3.6
12:48:29 4.9 3.7 91.4 87.2 0 0.0 34 2.9 2176 21.4 2158 11.8 3.7
12:48:39 4.9 3.7 91.4 88.0 0 0.0 18 2.6 2128 21.0 2118 11.8 3.7
12:48:49 5.6 4.5 89.9 88.1 0 0.0 35 2.7 2097 21.2 2082 11.7 3.7
12:48:59 6.8 4.6 88.7 88.0 0 0.0 60 3.1 2204 21.4 2228 12.1 4.0
12:49:09 5.6 4.0 90.4 88.0 0 0.0 15 2.9 2091 21.3 2072 11.7 3.6
12:49:20 13.9 4.9 81.2 87.6 0 0.0 14 2.7 2103 21.2 2084 11.6 3.7
12:49:30 9.1 4.0 86.9 87.5 1 0.1 70 2.8 2111 21.1 2129 12.1 3.8
12:49:40 7.4 4.8 87.8 87.3 1 0.0 52 4.5 2195 21.4 2200 12.0 4.1
</pre>

Two servers (serverSocket) behind the SAME firewal !?l

Hello !
I dont know if this is the right forum for this qusetion,
and maybe the question is really stupid, but ...
I made a little client/server application, just to
play around with sockets and serverSockets.
It works fine, but now i ask myself:
What happens, if there are two of my servers running on two
different computers in the SAME LAN behind the SAME
firewall/internetgateway, and both are listening on the SAME
port ?
F.e. if the extern IP of this LAN/(its gateway) is
204.556.234.123, and a client in the internet is connecting
to it on the port, on which the two servers behind the firewall
are listening, WHO'S ANSWERING ?
(If this port is set to "open and forward" in the firewall)
The one with the shorter patch-cable ? ;)
How is it possible for the client to differenciate
this two servers in that LAN?
The only logical solution i found is that this two servers
have to listen on different ports,
but i think there has to be another explanation and/or solution.
Do i have to take care about situations like that in my
server-application ?
I have the feeling that i have to...somehow.
OK, you see i dont know much about this...
i would be very thankful for every hint and explanation.
Thank you very much,
greetings,
huni.

F.e. if the extern IP of this LAN/(its gateway) is
204.556.234.123, and a client in the internet is
connecting
to it on the port, on which the two servers behind the
firewall
are listening, WHO'S ANSWERING ?
(If this port is set to "open and forward" in the
firewall)Whichever one the firewall is told to forward it to!
The two computers running your server have unique addresses on the internal network. The firewall will forward incoming connections to one of those addresses. Maybe it can do some simple "load balancing" by forwarding some connections to one server and some to the other, but still, any particular connection will only go to one server.

How to connect to two servers behind a firewall

Let's say I have a LAN with two independant servers, one at 192.168.2.10 and one at 192.168.2.20
As it is now, I can fwd the 548 port to one server and it works.
Is there a way i could connect to both servers, maybe having one and port 548 and one at 549, so if I connect using the default address, it will connect to server A and to the adress plus the port :549 it will connect to server B
I tried forwarding port 548 to one server and port 549 but for now, my router/firewall complains that I have a port conflict.
What's the simplest way to achieve what I would like.

Oh man, I think I've seen the answer to that in these treads, but it'd be around 20-30 pages back... iirc it would involve editing Hosts' or hppd file... BIG Help I know... but perhaps a real expert will be along shortly!

Two servers, one RAID

Similar Messages

Maybe you are looking for