After we installed the TECTIA Server SSH version 18.104.22.168 on a Windows 2003 server and reboot it, the server will be corrupted.
It seems to be happened on hardware related.
Because some one tells that there is a limitation on number of CPU in a server for TECTIA supporting. The maximum CPU will be 30 only, if over this number, TECTIA cannot support it and will appear problem. Unacceptable!
Our new server is Dell R910 with 4 X Eight cores = total 32 CPU.
Any expert have this information and suggestion to let us solve this issue ASAP ?
Actually, it looks to be Microsoft problem, Microsoft recognized the problem in:
and they provide a hotfix for it. The customer should download/request this hotfix from Microsoft.
To summarize: The problem occurs only in this scenario:
When the Lasss.exe process initializes, it creates some heaps to improve heap handling performance. If the number of processors is 32 or more, and if one or more third-party authentication packages are installed, the Lsass.exe process needs more locked pages. Therefore, the issue that is mentioned in the "Symptoms" section occurs when the minimum working set size is reached.
answered Jul 12 '11 at 15:31
Now that you mention the 32 CPU's I recall another case not too long ago that had a similar problem.
Unfortunately I'm sorry to say there is no solution at the moment for this. The problem lies in an underlying component of Tectia Server (sshdap) which is used in performing user authentication on Windows environments. For some reason on host with this large number of CPU's (30 or more) this component just does not work properly.
The only solution is to replace this component with other alternative methods for performing user authentication. We have opened a case for this on our backlog:
However doing this change requires a considerable amount of work which is not currently planned for an upcoming release. Even though this indeed quite a critical issue when running the server in this type of hardware (which in fact makes the server basically unusable in this kind of setup) it is not a common scenario among our customers. You are the 2nd customer (that we know about) that has run into this.
Given that we only have a very limited set of resources to distribute between release new versions and fix problems on released products we need to prioritize fixing the issues which affect a large set of our customers. This will definitely be fixed in a future version of Tectia Server, but just not likely to happen any time soon.
Since the Server is basically unusable in this type of environment you should be entitled to a full refund. Please contact our support, your local distributor or our regional sales for proceeding with a refund.
answered Feb 24 '11 at 21:02
Apparently there is a workaround which involves starting Windows and limiting it to use only 30 CPUs. However I agree that this is not that useful to just not be able to use the full resources on the system and this bug should just be fixed on Tectia Server.
This is done by adding the following to the boot.ini:
answered Feb 25 '11 at 09:17