UNIX/NT Interoperability

sbeavers · Post by **sbeavers** » Tue Mar 06, 2001 10:54 pm

We have texis running on both UNIX(64 bit) and NT systems. Can I create a database on one machine ( the NT box ) and copy it via FTP to the UNIX machine, or vice-versa, and have it all hold together?

sabety · Post by **sabety** » Wed Mar 07, 2001 6:08 am

I have similar issues. Trying to go from NT to Digital unix, but no cpdb utility on NT only copydb. I tried running cpdb from digital like this

cpdb -d /usr/netscape/server4/docs/webinator/db -h myhost -r myhost/db -t mytable.tbl -g

but error message resulted "connection refused in the function ezclientsock/could not connect"

is cpdb running on some other port number? or is the error indicative of something else?

I am able to see the remote data directory when browsing in netscape and even tried running the NT texis daemon (even though docs say don't need it), but still no help.

thanks

sabety · Post by **sabety** » Wed Mar 07, 2001 10:15 am

I have the following version:

Commercial Version 3.01.977332554 of Dec 20, 2000

sabety · Post by **sabety** » Thu Mar 08, 2001 11:22 am

excellent! works both ways. But, when pulling tables up from the digital machine (with -g option), the NT process ended with this message after the data transfer was complete:

000 Internal logic error
000 ABEND exception 0xC0000005 (ACCESS_VIOLATION)

should I be concerned about the integrity of the data in the new database on the alpha?

sabety · Post by **sabety** » Fri Aug 24, 2001 6:03 am

On some largish database tables I am getting an error on the NT end of things as follows:

read: Bad file descriptor
fillbuf: Bad file descriptor

I checked the disk space on both machines and there is plenty.

I noticed a pattern however. On the alpha server, cpdb stops at the same point each time and I get back to the command prompt. On the NT side, cpdb keeps running for a while before it gives the above error. The connection is obviously lost and NT can no longer see the host.

I ran a kdbfchk on the tables that copied unsuccessfully and no corruption was found. I also did a count(*) select to see if it was just the teardown but the number of records were much fewer than the source tables.

sabety · Post by **sabety** » Fri Aug 24, 2001 10:26 am

The tables are coming from an NT database, so they are under the 2GB limit. On one table of 7500 records, only 2500 were coming across. I did notice on the alpha that free memory dropped to zero as cpdb was doing its processing. I am using a DS10 with 256MB ram and 9 GB SCSI disk. The table above was about 350MB in size.

sabety · Post by **sabety** » Mon Aug 27, 2001 3:50 am

Here is the messsage for the version.

Texis Web Script (Vortex) Copyright (c) 1996-2000 Thunderstone - EPI, Inc.
Commercial Version 3.01.971289085 of Oct 11, 2000 (alpha-dec-osf4.0)

If this doesn't work, I suppose I could write two scripts for pulling the raw data across.

foosh101 · Post by **foosh101** » Wed May 14, 2003 9:13 am

Whats this about a 2GB limit. Also, I know this is unrelated, but, is there a particular file system I should use on the drive I have texis running? How about the drive I have the table and/or indexes. Should I specify a certain allocation unit size?

foosh101 · Post by **foosh101** » Wed May 14, 2003 9:59 am

I'm currently running Windows 200 Server. Just wanted to make sure there was not a limitation?

Post by **mark** » Wed May 14, 2003 10:06 am

Run
texis -dump
Near the bottom it will say 32-bit or 64-bit files. 32 bit files are limited to 2gb.

You should use a robust filesystem such as ntfs, not fat or vfat. But Texis will work with whatever you have.