LYCN with SQL alwayson"three" follow-up questions

In the first two sections, we goose that lays, rescue, finally married mating (AlwaysOn), Toward ... Stop, no test, still.

Yes, we do not have a back-end high-availability test, how to measure?

In the case of a client connection, close a backend database node and see if the client responds.

Exchange 2010-Cut DAG nodes have to be disconnected (manually cut active and passive is not, you shut down a full active copy of the MBX try?) ), you have a grafted Lync off backend node What mercy how can the client be unresponsive? Give it a try.

We turn off the current master copy and watch the client's response, and see if the right-hand pane has a water check mark. Hehe ...

650) this.width=650; "Width=" 796 "height=" 602 "title=" Image65 "style=" border:0px;width:750px;height:582px; "alt=" Image65 "src=" http://s3.51cto.com/wyfs02/M00/6E/02/wKiom1VwjF7TrLRtAAO-81q9rZY539.jpg "border=" 0 "/>

No, not enough to see again ...

650) this.width=650; "height=" 226 "title=" image "style=" border:0px; "alt=" image "src=" http://s3.51cto.com/wyfs02/M01 /6d/fe/wkiol1vwjfuzsqzvaaecnditqms657.jpg "border=" 0 "/>

OK, back to the chase, we border the response of the Lync client ... That is ... No response ...

650) this.width=650; "Width=" 744 "height=" 437 "title=" image66 "style=" BORDER:0PX;WIDTH:711PX;HEIGHT:417PX; "alt=" Image66 "src=" http://s3.51cto.com/wyfs02/M01/6E/02/wKiom1VwjF7wEP6VAAGQtEgk2E4198.jpg "border=" 0 "/>

Open the log on the Lyncfe to see? Very unlikely! A bunch of errors.

Comfort yourself: It doesn't matter, exchange shuts down a MBX will also make a lot of mistakes ...

650) this.width=650; "height=", "title=" image68 "style=" border:0px; "alt=" image68 "src=" Http://s3.51cto.com/wyfs02 /m02/6d/fe/wkiol1vwjfzglpnnaall10mrf9c394.jpg "border=" 0 "/>

Cut back to the client, then look?

Sure enough, Tiu Renderen, people in the day to see, do not believe up to see heaven around who, do not die will not die!

650) this.width=650; "height=" 478 "title=" image69 "style=" border:0px; "alt=" image69 "src=" Http://s3.51cto.com/wyfs02 /m02/6e/02/wkiom1vwjf_gz9mraag26sn85tk978.jpg "border=" 0 "/>

Read the front-end log in detail, and find such two articles.

650) this.width=650; "height=" 471 "title=" 5 "style=" border:0px; "alt=" 5 "src=" http://s3.51cto.com/wyfs02/M00/6D/FE/ Wkiol1vwjf3xppxjaakfsp7ikku847.jpg "border=" 0 "/>

650) this.width=650, "height=" 473 "title=" style= "border:0px", "alt=" src= "http://s3.51cto.com/wyfs02/M00/6E/02" /wkiom1vwjgdtc-e7aag5vu0sgkg602.jpg "border=" 0 "/>

Analysis, at this point we shut down the LyncBE-1 is the primary replica node, then the AlwaysOn listener will send the request to LyncBE-2, in other words, the Lyncfe front-end, unable to connect to the Lync database on LyncBE-2.

Why is it? contoso\lyncfe$ This is a computer account ...

650) this.width=650; "height=" "title=" QQ20150605010428 "style=" border:0px; "alt=" QQ20150605010428 "src=" http:/ S3.51cto.com/wyfs02/m01/6d/fe/wkiol1vwjf2yybidaack7su764c115.jpg "border=" 0 "/>

You must be waking up now. Already, is the relationship of the SQL login name.

Let's compare the list of logins for the two back-end databases:

650) this.width=650; "height=" 292 "title=" Image70 "style=" border:0px; "alt=" Image70 "src=" Http://s3.51cto.com/wyfs02 /m01/6e/02/wkiom1vwjgdqwgb-aaeggbrmw3w499.jpg "border=" 0 "/>

650) this.width=650; "height=" 261 "title=" Image71 "style=" border:0px; "alt=" Image71 "src=" Http://s3.51cto.com/wyfs02 /m02/6d/fe/wkiol1vwjf2hfr8baaeckfzs89u640.jpg "border=" 0 "/>

That is, when the topology was first published, LYNCFE created the database on LyncBE-1 and added the Lync Service Account group to the SQL login and assigned the login role, then we did AlwaysOn synchronization, and only the database was synchronized. And so important login name (5 functional groups!) ）！ We're not syncing!

In other words, we need to manually add a login name for some of the functional accounts of Lync on the LyncBE-2 node.

It is very simple to operate because there are LyncBE-1 nodes that can be referenced, and we know where to configure and what permissions to use.

As I have already done, these logins are configured with the same permissions as "Connect to SQL", so we just need to add a global universal group inside the domain, drag these Lync feature groups in, and then create a login name for this global universal group on LyncBE-2. and assign the LyncBE-2 connect SQL permission.

If you have a train of thought, do it! ：

650) this.width=650; "height=" "title=" QQ20150605011243 "style=" border:0px; "alt=" QQ20150605011243 "src=" http ://s3.51cto.com/wyfs02/m02/6e/02/wkiom1vwjggt-dkxaadq0ozc804721.jpg "border=" 0 "/>

Add Member

650) this.width=650; "height=" 394 "title=" image74 "style=" border:0px; "alt=" Image74 "src=" Http://s3.51cto.com/wyfs02 /m00/6d/fe/wkiol1vwjf6ybc1xaah3l6knzjk818.jpg "border=" 0 "/>

Add complete

650) this.width=650; "height=" 595 "title=" Image75 "style=" border:0px; "alt=" Image75 "src=" Http://s3.51cto.com/wyfs02 /m00/6e/02/wkiom1vwjghzb4rsaafk9rp8hsq341.jpg "border=" 0 "/>

Then open the SQL console above LyncBE-2. Right-click Security-Login-New login name

650) this.width=650; "height=" 278 "title=" 1231 "style=" border:0px; "alt=" 1231 "src=" http://s3.51cto.com/wyfs02/M01/ 6d/fe/wkiol1vwjf6xvswxaaetfgxloiu093.jpg "border=" 0 "/>

Click Search,

650) this.width=650; "height=" 473 "title=" image "style=" border:0px; "alt=" image "src=" http://s3.51cto.com/wyfs02/M01 /6e/02/wkiom1vwjgljvmglaagnwyp4z1o046.jpg "border=" 0 "/>

Note that the default is not to check the group, or the default is not allowed to add the group in. We need to tick and then enter the group name Lyncbelogin.

650) this.width=650; "height=" 445 "title=" image76 "style=" border:0px; "alt=" image76 "src=" Http://s3.51cto.com/wyfs02 /m02/6d/fe/wkiol1vwjf_gukz2aahrnln6dee109.jpg "border=" 0 "/>

Then click on the security object on the left, click Search, select Server LyncBE-2

650) this.width=650; "height=" 482 "title=" Image77 "style=" border:0px; "alt=" Image77 "src=" Http://s3.51cto.com/wyfs02 /m02/6e/02/wkiom1vwjgkj2h4jaaiimo57zmi669.jpg "border=" 0 "/>

In the following permissions, tick "connect SQL"

650) this.width=650; "height=" 549 "title=" image78 "style=" border:0px; "alt=" image78 "src=" Http://s3.51cto.com/wyfs02 /m00/6d/fe/wkiol1vwjf-hmgozaahrdtf35wm890.jpg "border=" 0 "/>

Then click OK so that you can have the Lync Service Group connect to the service account LyncBE-2!

650) this.width=650; "height=" 372 "title=" 3333 "style=" border:0px; "alt=" 3333 "src=" http://s3.51cto.com/wyfs02/M00/ 6e/02/wkiom1vwjgpz45r6aaf3ouukapk562.jpg "border=" 0 "/>

In fact, when the operation to this step, as long as the added success, the Lync client side will immediately have a response, that is no longer prompt during the interruption limited functionality available.

But I didn't leave that ...

OK, next we'll start LyncBE-1, and we'll try to turn off two backend nodes. Observe the client response at the same time.

No response.

650) this.width=650; "Width=" 793 "height=" 451 "title=" Image80 "style=" border:0px;width:725px;height:410px; "alt=" Image80 "src=" http://s3.51cto.com/wyfs02/M01/6D/FE/wKioL1VwjgCS-QXOAAHLIGYqufc967.jpg "border=" 0 "/>

Still no response ...

650) this.width=650; "Width=" 796 "height=" 437 "title=" image79 "style=" border:0px;width:725px;height:401px; "alt=" Image79 "src=" http://s3.51cto.com/wyfs02/M01/6E/02/wKiom1VwjGOyUGbuAAGTNLDKC3U231.jpg "border=" 0 "/>

I didn't even have an error in the front-end log!

650) this.width=650; "Width=" 823 "height=" 391 "title=" image81 "style=" border:0px;width:735px;height:345px; "alt=" Image81 "src=" http://s3.51cto.com/wyfs02/M02/6D/FE/wKioL1VwjgGBKn9HAAN2HTFoZEY234.jpg "border=" 0 "/>

It has been ... Basically, in terms of connectivity, this architecture is permissible and reasonable.

When a backend node fails over, the client has no sense at all. But functional aspects, until the press, I tested the basic IM functionality, there is no problem ...

As for other components, such as archive monitoring ... I'm not going to say it.

Cdr...... Yes, this library is created when you create the front end of the installation ... When the topology is published, it has nothing to do with it, so this 56202 error will only let him down.

At present, the solution I think of is to find a normal Lync 2013 environment, record the configuration of the database, such as the path, the initial size, and then the LCSCDR this database is manually established on the current schema, and then added to the AlwaysOn availability Group. As for the operation, leave it to the people who are interested ...

650) this.width=650; "height=" 480 "title=" Image64 "style=" border:0px; "alt=" Image64 "src=" Http://s3.51cto.com/wyfs02 /m00/6e/02/wkiom1vwjgsxg1gjaaghb2bdzni449.jpg "border=" 0 "/>

So, this architecture is still flawed and risky. Although I have only found this problem at present, but after all, is the test environment, the rest of the components of the lack of persuasion ... If Lync has a systematic diagnostic tool, it can do a health test or stress test, if you crossing found other problems, also welcome message exchange. Although Heterodoxy, but in the small and medium-sized environment, the database is mostly stacked in a scene, the existence of this architecture is a very cost-efficient high-availability solution!

This article is from the "Castamere Rainy season" blog, be sure to keep this source http://sodaxu.blog.51cto.com/8850288/1658696

LYCN with SQL alwayson"three" follow-up questions