I do my own databases; have written marvelous scripts for automated database
all broken in one way or another by the "improvements" in the directory
and all the rest that you have so marvelously documented.
It may take a lifetime to learn, but
IF IT AIN'T BROKE, DON'T FIX IT!
Thanks for warning me about case sensitivity -- that might explain some
Telephone: 919-316-4569 (NIEHS: 6-4569)
Scientific Computing Support Group
NIEHS Information Technology Support Services Contract
(Science Task Monitor: Jack L. Field( field1 at niehs.nih.gov )
National Institute of Environmental Health Sciences
National Institutes of Health
Research Triangle Park, North Carolina
> From: Michael Black
> Sent: Wednesday, September 7, 2005 7:20 PM
> To: info-gcg at magpie.bio.indiana.edu> Subject: [Info-gcg] GCG 11.0
>> Is anyone else frustrated with GCG 11.0 and Seqweb 3.0?
>> My first peeve (which I've communicated to Accelrys) was the lack of
> printed documentation - not the program docs, but the admin docs. Those
> should have been provided in printed format.
>> Second, I learned today, that dbnamemap.conf is now case sensitive,
> which is not documented, and its predecessor never was. An example, we
> have the C. elegans database locally - I named the sections "worm_I",
> "worm_II", and so forth. But 11.0 only works with them if the
> chromosome roman numbers are lower case - which was never the case with
> GCG 10.0 through 10.3.
>> Another example - admins of GCG 10 are familiar with the gcgdata
> directory (which in our case held sym links to our database
> directories). In 11.0, there is a "db" directory, which is blank - no
> documentation on what it's there for, why it exists, nothing (should it
> hold sym links to the database files - ours are on a RAID?). What the
> heck is it for, since .conf files seem to cover all that now?
>> GCGSRS appears to need a ".wp" directory and files (for gcgadmin) to
> work correctly (that was the case for the patch for 10.3, which I only
> learned about 4 months after it was released), but this directory is not
> created by default with the install of II.0?
>> We are still running GCG 10.3 for our users, and have CAREFULLY followed
> the install instructions for 11.0, and running it at the same time.
> Yet, when I invoke GCGSUPPORt for 11.0, I still have to move into "sbin"
> and run utils directly from there - sbin does not seem to be added to
> paths when GCG 11.0 is sourced in at startup?
>> We have not even gotten into SeqWeb 3.0, but will shortly.
>> So far the transition has not gone smoothly. I've had to futz with a
> lot of .conf files, discovering things about their syntax that is not
>> Has anyone had a smooth transition from 10.3 to 11.0? We run GCG on a
> SunFire 480 (dual cpu's, 10GB RAM - databases are on an Apple XRAID -
> about 1.98TB total, sliced into two ~917+GB wedges). We do not
> subscribe to any Accelry's database service - do that ourselves (PERL
> scripts), and do have a lot of custom db's (ie. not just the regular
> NCBI ones, nor just UniProt). So our db conf files are non-standard.
> But, still, simple things like case sensitivity aren't even documented,
> after years of not being so. This irritates me.
>> Just looking for observations/comments about 11.0 from admins.
>> Cheers, Michael
> Michael B. Black, Ph.D.
> Molecular Biology Computing Support
> University of Virginia
> ITC-Academic Computing Health Sciences
> Info-gcg mailing list
>Info-gcg at net.bio.net>http://www.bio.net/biomail/listinfo/info-gcg>>