Hi

Currently running BES 5.02 with HA. Lotus Domino 8.5.2

BAS is installed on both the primary and secondary servers and the services are running on both.

I have a pool name setup and both the BES configs are pointing to this single pool name

For some strange reason, I can only log in to the BAS when I shut down the BAS service on one server. With both BAS services, login fails. Would this suggest an issue with the pool name

How are you logging into the BAS, are you logging in via notes ID or through the BAS itself?

__________________
MCSA 2003 +M MCITP | EMA
Who cares, RIM will just change how they do Cert's again

I had this prob initially try this:

In BAS go to:
BlackBerry Solution topology > BlackBerry Domain > Component view > View (BlackBerry Administration Service)
& edit the BAS. Configure the following as below:

Transport protocol configuration
BlackBerry Administration Service pool: TCP with TCP PING

In my environment I'm runnign on Hyper-V VMs.. and the virtual switches did not like UDP Ping and I had the same prob as you. Switching to TCP made everything much better

bhorigan - im logging in using Notes credentials. I have checked the DIIOP services and these are running.

Zippygit - Yep tried that also and then found I could no longer access the BAS, fortunately it was a dev system so ended up doing a clean install of the BES.

Yesterday I spoke to a guy in our carriers support team, who said only one instance of BAS should be running at any given time and that a manual failover would be required. Kind of find this an odd concept as it does mean that if the main server dies and takes the BAS out also. Whilst the other services should failover, I will have to login to the system and manually start the BAS services. Not ideal in my opinion.

Wonder if this is actually the case or are people able to have multiple BAS services running which access the same Database?

ONly run one BAS at the same time... that's a load of horse hockey. I have 4 node BAS up and running, and it works reallu well. In my situation I have a specific startup "quirk" where if don't do the below BAS will not start/converge properly.

1 BAS server needs to successfully/fully start first and assume a BAS pool singleton role if this does not occur the cluster will no converge and will not work as you expect. Here's what I do:

Stop all BAS services by stopping the Blackberry Administration Service - Native Code Container Service (BASNCC), this will stop all BAS Dependant services.

Start The BlackBerry Administration Service - Application Server (BASAS)one one node (this will start all BAS Dependant services). Then do the following on ONE NODE.

Start Task manager and view processes for all users. Look for a BAS-AS.exe process... and wait for it "warm up" in my case it has to get to ~530MB of RAM usage before is starts servicing user requests. Try logging in , if you can then start BAS on your second node and one it warms up you should have a 2 node BAS Pool.

In my specific instance (which I'm attributing to running BES in a hyper-v VM) I can only start one BAS node at a time, but once they're all up I'm good.

Hope this helps.

Update on this folks

Seems it was relating to the DIIOP process and the firewall.

within the BB Server config I have the name of the primary bes server, I found that if I happened to hit the secondary server on login, we have netscaler in place, this then failed due to it being unable to communicate with the DIIOP process on the primary.

I have now opened up the firewall on the DIIOP port and this now works.

Many thanks for the above advice, especially the "warm up" that solved another issue for me.

Of course the problem I will now have is on failover as I guess the BB server config will not change to the other server. So figure I now need to sort out a DNS name for the DIIOP that points to both these servers.

Another update on this one - seems this was not the case and I still have the same problem, although have further information.

It seems the issue relates to what ever BAS server happens to be shown as the Partition Singleton (highlighted in red)

I have switched the services so as to change which of the two servers running BAS and can replicate this.

Unable to find any information on where to go next.

Wonder if anyone has any ideas on this.

Thanks in advance

Note in my setup on running on exchange so there are more than likely some differences in implementation but here's hoping this is of some help.

I had somthing similar when runnig BES 5 SP1, in SP1 seting up TCP Ping was a royal pain in the arse. I opened my case with RIM and reviewed my config and I had a typo in the config files, once addressed all was well. SP2 introduced a GUI for setting TCP Ping, so if you're not running SP2 I woudl strongly reccomend upgrading & letting it implement TCP ping properly.

Apart from that I the only other things I can think of are:
make sure you read the install prerequisites and implemented properly, if you dont' you'll have simialr issues (E.G. If IPV6 is enabled it will cause you tonnes of pain).

Consider doing a full BES shutdown (power off both nodes) let one warm up fully then power on the seconffand let it do the same (after you've confirmed install prerequisites of course).

If you're still getting partition singleton issues you will really want to consider contacting RIM (or your carrier if they have BES support - Rogers in canada offers free bes support for their customers, not sure about carriers the US) support to get them to take a look at your setup as somtimes lots of little things can get in the way.

Hope this helps and good luck.

Z

OK thanks for the update - I have 5.02 installed so figure that is SP2. I did try changing the UDP Multicast to TCPPING but this ended up stopping all access to the BAS so I had to do a complete re install of the BES.

I have a support call out with my carrier but since they initially said only one BAS I thought I would check to see what other folk are doing.

Thanks again for your help. Will keep the forum posted.

OK latest update - KB25584

Other Blackberry Administration Service nodes appear as Partial Singletons and are unable to communicate with them.

My problem is not other BAS service nodes appearing as Partial Singletons, but this is the info sent from the carrier and thought it was worth posting for information.

Essentially what I have just done is a total power off as Zippygit suggested, thanks for that.
What we had been doing was simply shutting down the service which clearly was not sufficient.

I left the primary server on for just over an hour before powering on the second server. This seems to have done the trick.

Thanks all