[[https://www.revelation.com/|Sign up on the Revelation Software website to have access to the most current content, and to be able to ask questions and get answers from the Revelation community]]
==== Network Protocol (AREV Specific) ====
=== At 27 JUL 2000 02:02:56PM Chris wrote: ===
{{tag>"AREV Specific"}}
I'm going to ask this question, but it may not be relevant to AREV specifically.
I have recently started maintaining a Novell network that is running a program that uses AREV as its backend. The program was running along fine until a change was made to the network protocol. It was running with "All Networks" and was changed to "IPX". It had been running for 4 years as "All Networks" and never had a problem. Suddenly one of the tables has become corrupt and data has been lost. It was changed back the following day, because there were several fatal errors and other unusual instances. It happened again a couple weeks later when another technician made the same change. Again, several fatal errors and severe data loss.
My question is this... can a change like this cause such severe problems and corruption? Should the Indexing station been shut down and everyone been made to log off before such a change? Or is this a coincidence, and it would have happened anyway?
I know my questions may be diffcult to ask. I am not familiar with AREV, and I'm trying to expand my understanding of it.
Thank you,
Chris
----
=== At 27 JUL 2000 02:17PM Don Miller - C3 Inc. wrote: ===
Chris ..
]]My question is this... can a change like this cause such severe problems
and corruption? Should the Indexing station been shut down and everyone
been made to log off before such a change? Or is this a coincidence, and
it would have happened anyway?
I know my questions may be diffcult to ask. I am not familiar with AREV,
and I'm trying to expand my understanding of it.
ABSOLUTELY ...
The All-Networks driver is a variant of the MS-NETWORKS Byte-Range driver. Locks on the server are set in a file named REVLOCKS in every subdirectory containg REVnnnnn.LK / .OV files. The IPX / SPX Novell) driver sets locks via semaphore flags maintained in server memory. If you changed the driver while ANY workstation had files opened using the previous driver (particularly a dedicated indexing machine), file corruption is assured. Additionally, make sure that the workstation Frame Type is set correctly if you are using the Netware driver. You didn't mention if you were using the NLM IPX/SPX driver. If this is the case, then the All-Networks driver on the indexing machine will definitely have trashed the files since its I/O was totally bypassing the NLM's control.
Maybe this background will be useful
Don Miller
C3 Inc.
----
=== At 27 JUL 2000 04:25PM Steve Smith wrote: ===
Chris,
Don's comments are 100% reliable.
In any database situation where files are shared by multiple users across a network, there has to be a technique for ensuring files are locked and unlocked when in use. This means each user gets exclusive access to the area of the file with the record(s) they require. Once you change AREV's network driver, the locking doesn't work any more, and two users may update the file at once. Because records may span multiple files in AREV, and because they are layed out contiguously, any addition or adjustment of data without locking is going to cause damage.
steve@state-of-the-art.com.au
----
=== At 27 JUL 2000 05:01PM Kenny G wrote: ===
This is like letting two hippos into the bathtub at the same time. When it's all over with and things are back under control, there's not going to be much water left in the tub.
[[https://www.revelation.com/revweb/oecgi4p.php/O4W_HANDOFF?DESTN=O4W_RUN_FORM&INQID=NONWORKS_READ&SUMMARY=1&KEY=0C6F04FC16D8CA14852569290063258E|View this thread on the forum...]]