Forum Discussion
Setup for UAG proxying Activesync traffic
Hi all,
I was wondering if anyone has setup an F5 VIP going to a UAG which proxies the traffic to exchange. Below is a crude diagram of the setup.
Internet--->External F5--->UAG Array (UAG1 and UAG2)--->Internal F5--->Exchange
The external F5 load balancer VIPs between the 2 servers in the UAG array. The internal F5 VIP appears to be running correctly. I see the following errors 24 and 47 on the UAGs with this setup.
Error 24
The request from user xxxx at source IP address 1.1.1.1 to trunk 2010gaowa; Secure=1 failed because the request was unable to reply to an HTTP 401 request from application Exhcange 2010 EAS and OA of type ExchangePub2010. The session ID is C6B5DA33-51FE-4A7F-8D08-EA0B26850781.
Error 47
A request for application Internal Site of type InternalSite on trunk 2010gaowa; Secure=1 failed because a POST action without a content-type header is not allowed. The URL is /InternalSite/InitParams.aspx?referrer=/InternalSite/InternalError.asp&site%5Fname=2010gaowa&secure=1&error%5Fcode=17&policy%5Fid=&ErrorType=HTTPErrorCode. The source IP address is 1.1.1.1. The user is xxxx.
I want to know if someone has some suggestions or documentation on how to properly set up the external F5 to load balance the UAGs correctly. I've reviewed the F5 TMG guide and F5 UAG DA guide. Those documents don't seem to reference the setup I have.
Thanks in advance.
Bill
6 Replies
- What_Lies_Bene1
Cirrostratus
Hey Bill, I'm unlikely to be able to help here but I'm sure it'll save a bit of time for others if you could let us know what version of TMOS you are running and provide the Virtual Server, Pool, Persistence and any other applied profile configurations, ideally in tmsh format;
-[tmsh] list ltm virtual 'name'
-[tmsh] list ltm pool 'name'
-[tmsh] list ltm persistence ...
-[tmsh] list ltm profile http ... - wng_98840
Nimbostratus
Hi Steve,
Here is the config and version listed below.
BIG-IP 10.2.2 Build 852.0 Hotfix HF1
ltm virtual aaa.bbb.com-uag-https-vs {
destination 1.1.1.1:https
ip-protocol tcp
mask 255.255.255.255
persist {
source_addr {
default yes
}
}
pool uag01-uag02-aaa.bbb.com-https
profiles {
bbb.com {
context clientside
}
http { }
serverssl {
context serverside
}
tcp { }
}
snat automap
vlans {
prod-vip-amber
}
vlans-enabled
}
ltm pool uag01-uag02-aaa.bbb.com-https {
load-balancing-mode least-connections-member
members {
2.2.2.2:https {
session monitor-enabled
}
2.2.2.3:https {
session monitor-enabled
}
}
monitor uag-https-healthcheck
}
ltm persistence global-settings { }
ltm profile http http {
adaptive-parsing enabled
basic-auth-realm none
compress disabled
compress-allow-http-10 disabled
compress-browser-workarounds disabled
compress-buffer-size 4096
compress-content-type-exclude none
compress-content-type-include { text/ "application/(xml|x-javascript)" }
compress-cpu-saver enabled
compress-gzip-level 1
compress-gzip-memory-level 8k
compress-gzip-window-size 16k
compress-keep-accept-encoding disabled
compress-method-prefer gzip
compress-min-size 1024
compress-uri-exclude none
compress-uri-include none
compress-vary-header enabled
lws-width 80
max-header-size 32768
oneconnect-transformations enabled
pipelining enabled
ramcache disabled
ramcache-aging-rate 9
ramcache-cache-control-mode all
ramcache-insert-age-header enabled
ramcache-max-age 3600
ramcache-max-entries 10000
ramcache-object-max-size 50000
ramcache-object-min-size 500
ramcache-size 100
ramcache-uri-exclude none
ramcache-uri-include none
ramcache-uri-pinned none
response-chunking selective
} - What_Lies_Bene1
Cirrostratus
OK, thank you. I'd say in general it mostly looks OK. However, assuming NTLM authentication is being used then I don't think you should be using OneConnect as the two are not compatible prior to v11.
Also, nothing to do with your issue but you don't seem to be doing compression. Are you doing it on the internal F5's? - wng_98840
Nimbostratus
Steve,
I'll have to verify the authentication we are using with activesync. If we are using NTLM, should I disable the OneConnect?
I don't think we are using compression on the internal F5s. As best practice should we do compression?
Thanks,
Bill - What_Lies_Bene1
Cirrostratus
Yes, if using NTLM authentication, don't use OneConnect as well. Note this applies where a single layer of F5s are used. In your case things may be a bit more 'nuanced' but I'd certainly disable it all round and see if this solves the first issue. If not, put it back.
Compression is definitely a good idea for any HTTP traffic and will reduce bandwidth usage and improve apparent client performance significantly. I'd recommend it, but implement it only on one set of F5's. I'd say the external ones to prevent any possible issues with the proxies not understanding the compressed content. I typically configure deflate (rather than gzip) and a level of 8. - JoeTheFifth
Altostratus
a bit too late but I had the same error coming from a uag array and it solved with setting uo a cookie persistence profile. the problem here is I think packets are lost between the vip and the uag arrays servers...
Help guide the future of your DevCentral Community!
What tools do you use to collaborate? (1min - anonymous)Recent Discussions
Related Content
* Getting Started on DevCentral
* Community Guidelines
* Community Terms of Use / EULA
* Community Ranking Explained
* Community Resources
* Contact the DevCentral Team
* Update MFA on account.f5.com
