I saw a posting on the discussion group recently where someone was actually running Arev on the NT server machine. Be aware that the documentation specifically says that you should NOT do this.

AIUI, since the NT server is not ACTUALLY on the WAN, any access to the files on the NT Server would be be via direct access to the drive, and not via TCP/IP or the named pipes.

I suppose though, that if the REVPARAM file was set to SERVERONLY=TRUE that this should ensure that the NT service WAS being used.

Any feedback from anyone using the NT server to run Arev would be appreciated, since I have quite a few customers who would like to utilise the NT server for an Arev Index server.

Tim Sneller - mailto:[email protected]

"If it's not locking, don't come knocking….."

Indexing processeors are best run from workstations.

Steve

I experimented with this myself briefly. I was hoping to get the index updates to occour without slowing the network, and also to reduce the number of PC's in the server room.

If I remember correctly, you can achieve locking if you access the LH files through a network maped drive. This appeared to work well, but the experiment was stoped for other reasons (policical!).

I run ARev on an NT server, however this is a stand-alone development system only and only during testing is a workstation or two is ever connected.

1) Running ARev on the server pushes the CPU utilization up to 100% with NTVDM (ARev) taking up the majority of the CPU. This will significantly impact the perfermance of the server. This can sometimes be lowered by tinkering idle sensitivity.

2) NT server sometimes allocates huge blocks of resources to run ARev (I've seen NTVDM taking up 32MB of memory). Again, this can severly impact the server.

I would not recommend running ARev on the server in a live environment.

View this thread on the forum...