[tanteilisa]

I've just upgraded the bes from 5.0.0 to 5.0.1.

Everything seemed to run fine during installation. All of the services were started, except when I go to the BES logs, I find that the BBAS_AS, MDSS-AC and BBAS_NC logs were missing. I couldn't connected to the BAS webconsole through the browser either.

I found this in the BES\jboss\ejb\BAS-NCC_err.log:

Error encountered while starting server: null
java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Nativ e Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(Unknow n Source)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(Un known Source)
at java.lang.reflect.Method.invoke(Unknown Source)
at com.rim.bes.bas.system.ServerMain.main(ServerMain. java:662)
Caused by: java.lang.NumberFormatException: For input string: "p"
at java.lang.NumberFormatException.forInputString(Unk nown Source)
at java.lang.Integer.parseInt(Unknown Source)
at java.math.BigInteger.<init>(Unknown Source)
at com.rim.bes.bas.system.Encoder.decode(Encoder.java :206)
at com.rim.bes.bas.system.ServerMain.getKeyStoreRegis tryKeyValues(ServerMain.java:999)
at com.rim.bes.bas.system.ServerMain.initialize(Serve rMain.java:751)
... 5 more


I tried restarting all of the services in order and nothing helps.

Does anyone know what's going on?

[hugheser]

I'm experiencing the same thing after upgrading to SP1 yesterday. One of the servers is working OK but the other 3 are not. I'll update here if I discover anything.

[hugheser]

Here's a quick update. I haven't been able to find a solution to this so I created a support ticket with RIM. It has been about 3 hours without any contact but hopefully they will get back with me soon.

[besadmin23]

Do you have multiple IP NIC's on the box and if so, do you have more than one of them with IP addresses? If so, this may be the issue. KB18190***

You need to check the bind order of your NIC's. You need to ensure that the primary NIC is in the top of the Bind order.

This is the W2K3 instructions...
Right-click My Network Places and select Properties or Control Panel, and then click Network Connections.
Click Advanced, and then click Advanced Settings.
Under the Adapters and Bindings tab, verify that the correct binding order appears under Connections.

[hugheser]

No they have a single NIC. RIM never responded to my ticket so hopefully I hear something early tomorrow.
Thanks

[besadmin23]

ok, are you having the same problem as the first post? "I couldn't connected to the BAS webconsole through the browser either."

If so, how do you know one is working and the others not? Or are you just talking about logging not working? And to confirm, you have BAS installed on all three of the machines? Are you pointed to a remote machine for Attachment services? Unless you specifically installed BAS, I don't think you will see the logs on the machines it isn't installed on.

Just some random thoughts.

[hugheser]

Yes I am. I know it isn't working because I cannot access it by browsing directly to the machine's address. For example, https://machine1/webconsole/login does not work but https://machine2/webconsole/login does. I also cannot telnet to port 443 on machine 1 but I can on machine2. The logs for the BAS on machine1 stop at the moment I installed SP1 and haven't been back since. The BAS logs on machine2 went from maybe 20MB a day to several GB a day.

We use the BRK to handle account provisioning as well as user lists and such. If I point BESUseradminclient at machine2, it works for basic stuff but not for users lists. It times out. It doesn't work at all if I point it at any of the other machines. Lastly, machine2's CPU is pretty much constantly pegged with BAS-AS.exe causing it. This is all after the install of SP1 and MR1. None of these problems existed before then.

BAS is installed on all 4 machines. Two are primary with the other two as standby. This is confirmed by checking the BAS services on all of the machines.

I just finally got an e-mail from RIM telling me to call them so we'll see how that goes.

[besadmin23]

Originally Posted by hugheser

BAS is installed on all 4 machines. Two are primary with the other two as standby. This is confirmed by checking the BAS services on all of the machines.

Need to make sure we are talking about the right thing here. BAS can not be in either standby or primary mode. BAS is always on if it is installed. BES on the other hand can be installed in High Availability mode which will then allow the different nodes (physical or virtual machines) to be in either standby or primary.

BAS is the BB Administration Service. This is the web admin console.

Sorry if you already know this, but I am just trying to understand what you have going on better to help.

[hugheser]

Yes I am talking about the BlackBerry Administration Service. We have 4 BES servers. 2 primary and 2 standby. The BAS service is what isn't working. It is enabled on all of the servers but can only be accessed on one.

[besadmin23]

Have you configured a round robin pool name in DNS for the machines? If so, does an nslookup from each machine resolve each of the 4 addresses?

On the primary standby subject. Think of BAS as completely separate from the BES... even though they are installed on the same box. When a BES is in either Primary or Standby, the BAS isn't affected by this. It is a speparte entity all together.

Just to be sure on another thing. I know you said that each machine has only one nic. If these are servers, I can't imagine they only have one nic. Just to be sure, can you check the bind order on the NIC on each machine. Even if you have multiple but only one has an IP address, you should check. Unless other nic's are completely disabled, they will self assign a 169.254.x.x address and if your primary nic is lower in the bind order, you will have a prob with bas.

Go to your network control pannel and in the menus across the top, click on 'advanced' then 'advanced settings'. You will see the order there. Move the primary nic to the top and restart the BAS services on that machine.

[hugheser]

Yes they are setup with round robin in DNS but I am not using that name. I am browsing directly to the machine.

They do have only one NIC since they are virtual machines. Again, this all worked prior to installing the service pack on Sunday.

[hugheser]

I'm on the phone with RIM now and I was told it is a known issue with 5.0.1 MR1. They are working on a new MR. They are trying to come up with a temporary workaround at the moment but no luck so far. Hopefully they can figure something out.

[TimMunro]

Well I'm glad it's a "known issue". I've been experiencing the exact same problems with the BESUserAdminClient since SP1 MR1.

Interestingly I can query (list get stats etc.) and single user, but when I ask for all of them. I get a "time out".

Also, if I point the Client to the Stand by I get "Unauthorized" when using ad_auth, and "Internal error: Cannot find Dispatcher Instance:" when using BAS_Auth.

I really hope they get this fixed soon.

[hugheser]

Tim, I have the same problem with the BESUseradminclient. They said it is a known issue but they don't have a fix out yet. Nothing in the KB about it either. You might want to contact support to make sure you get the fix ASAP. They are supposed to call me back tomorrow afternoon with a workaround.

[besadmin23]

if they are vm's, are they all on the same physical host? or on separate hosts? Have you tried moving them all to the same host?

[TimMunro]

For the longest time they were on the same host. I've only just recently had them moved them to separate hosts. No change.

hugheser: If the work around works, please post.

Thanks.
--
Tim
BES 5.0.1/Exch 2007

[mahoward]

You cannot have BAS installed on the primary and secondary nodes of a cluster. I learned this the hard way as there is no documentation stating this. One instance claims status as the master and therefore the other instance cannot start.

I'm afraid you are going to be stuck with only 2 BAS instances in your failover topology instead of the expected 4 unless you build some more servers.

Here is my ticket resolution from RIM:

Hello,

Your Service Request has been logged and the following course of action has been
provided in response to your inquiry. If you have any questions concerning this request,
please reply to this email within the next three (3) business days.

Details:
TX3 BES 5.0 MR2 Exchange 2003 SP2 SQL 2005 Unable to login to BAS

Resolution:
Stopped BAS services on fail over BES

[mahoward]

Another issue: If you move BAS hosts from the same subnet to different subnets, you will have to change the heartbeat communication between them from UDP (default which doesn't traverse routers) to TCP. There is a KB article on doing this and it is a real pain.

[hugheser]

I don't think that is it in my case. It has worked since May and didn't stop working until the latest updates. Also the services are running on all 4 servers. That isn't the problem.
Here is a quick update. By undoing what the tech had me do yesterday, I was able to get the BAS running on both of the primary machines. I can login and do everything. The BAS-AS service however still pegs the CPU and the BESUserAdminClient fails to do any of the more complicated tasks like create a user status list. T Support should call this afternoon and hopefully we can get this solved.

[jah8903]

Any updates on this post as yet?