Need help in planning for High Availability on Microsoft


Warning: count(): Parameter must be an array or an object that implements Countable in /home/styllloz/public_html/qa-theme/donut-theme/qa-donut-layer.php on line 274
0 like 0 dislike
11 views
Sdraste. Came to the hands/feet/head to plan for high availability in a particular office (not a work I haven't replaced :)).

So, this:

1. The desire to eliminate critical points of failure and increase availability for the following services:
1.1. File server
1.2. Exchange
1.3. CRM+ERP-server
2. The need to replace some server hardware.
3. The presence of a roasted rooster that pecked too economical command of the company (cost of one hour downtime, but it seems that the perceived need exists).
4. The possibility of building a second server room in the building that will eliminate the potential risk of global catastrophe in a single server: fire / flood / meteorite etc.

Assume of building a Failover Cluster for Hyper-V, planning to virtualize:
* file server
* exchange with distributed roles (now physically the exchange server exists in a single copy)
* SQL server (erp).

Internet access, obviously, you will need to duplicate by using a Forefront TMG array, but it is not in the first place.

At the moment the task is cosmetici all this economy.

Issues — weight, really: will list their thoughts in response asks for optimization and advice from the elders, who ate oysters:

1. Microsoft recommends that you deploy a Failover cluster to use a highly available data store that is offering as a pre-deployment storage script implementing a separate cluster. In fact, it turns out that required the implementation of two independent clusters: (a) file server and (b) hyper-v, use resources (a). It turns out that requires at least four physical nodes, two per cluster. Given the possibility of placing the nodes in different server rooms, the approach is correct from the point of view of fault tolerance (and given the policy of licensing of an Enterprise edition of windows, like, and the ability to carry 16 virtual machines — please correct if wrong)
2. The existence of a separate cluster as a file server assumes the availability of fault-tolerant storage. Taking into account the different server, I suppose, it is advisable to deploy the correct NAS. Question: how to organize fault-tolerant NAS? On what equipment to watch, based on the desire to optimize the budget of the event?
3. Is it enough for the normal functioning of the two above-mentioned clusters of the existing network equipment? Now for switching servers exploit a stack of Gigabit AT-AR9424, the performance is satisfied. Whether you want organization more productive networks (optics?) or adequately enough available?
4. How to optimize the system from the point of view of the budget? What equipment is necessary to apply in the scenario described? Who sold a similar — please share your experience, please.
by | 11 views

2 Answers

0 like 0 dislike
Most of those services about which you have written, you can cluster at the level of services, such as domain controllers, DNS, DHCP, file servers (DFS), Exchange server (DAG), SQL (for example, Database Mirroring). In most cases this will provide a greater level of service availability (as it will not only protect from failure of iron), though may require a greater number of licenses (Exchange, SQL), and also will eliminate single point of failure in the shared storage; more effectively utilize server resources (rather than just reserving them under Failover Cluster); use DAS (sas/scsi storage drives or internal servers) as storage for virtual machines and application data, instead of shared storage.
\r
\r
exchange, diversity of role

Are you sure it needs to be done? Of course, depends on the load and number of users, but if <300 people is not worth it to bother too much, except, perhaps, dedicated front ends (for example Exchange Edge or other mail relay) for the anti-spam or anti-virus scanning of incoming messages.
\r
\r
In fact, it turns out that required the implementation of two independent clusters: (a) file server

NAS (CIFS or NFS) are not officially supported as storage virtual machines on Hyper-V, especially with DFS. If you expect otherwise cluster file storage (e.g. with the same Failover Cluster) you'll need shared block storage entity (DAS, SAN). Sense then a kitchen garden to fence?
\r
\r
(and given the policy of licensing of an Enterprise edition of windows, like, and the ability to carry 16 virtual machines — please correct if wrong)

Not really, when you use a Enterprise license you can run up to 4 OS instances on the same server to which this license is attached, if you have 2 Hyper-V servers, then you will be able to raise the 4-d VM on one and 4-d on the other.
\r
\r
Question: how to organize fault-tolerant NAS?

With regard to the correct NAS — as already mentioned — is not supported. If you want fault-tolerant storage (relatively cheap), look at Starwind or self-assembled Linux + ietd + drbd + heartbeat (but, of course, such a decision will not be supported).
\r
\r
Is it enough for the normal functioning of the two above-mentioned clusters of the existing network equipment?

50/50 either enough or not enough.
by
0 like 0 dislike
In your company plan to do the following:
10 servers is citrix xenserver.
Storage, connected to these same 10 servers for Vibram.
All services(sql, terminal, mail) are virtualized.
Accordingly, fault tolerance, and load balancing takes care of tsitriks.
\r
Storage alcazarquivir to make a budget. He SHD we go out in the area of 500 thousand, so that backup storage will obviously be expensive.
by

Related questions

0 like 0 dislike
3 answers
asked Mar 28, 2019 by isakov
0 like 0 dislike
2 answers
0 like 0 dislike
5 answers
0 like 0 dislike
2 answers
0 like 0 dislike
2 answers
asked May 11, 2019 by alex_terekhov
110,608 questions
257,186 answers
0 comments
27,902 users