[Zinc-fans] problem with Search and Browse utility

[John J. Irwin]

Hi Marc & ZINC-fans

Thanks for your feedback about slow ZINC queries.

We have had some major performance stalls in ZINC over the past few days
(Thanksgiving weekend).  We figured out what is going on with that, and
ZINC has been more stable this afternoon.

Another possible cause of slowness has been that I have been
aggressively loading and depleting (removing) molecules based on the
latest vendor catalogs. Sometimes I get a bit carried away with this,
and try to do too much at once, which slows everyone down.

Here is what we will do:

1. I will slow down with the database updates (4M molecules deleted, and
nearly as many new molecules were added during November). That should
significantly reduce the load on the database servers.

2. We will move up the addition of a new replicated database server that
will increase our ability to handle requests by 50%. (i.e we have 2 now,
we will have 3).

3. I will monitor the database and the website, and take advice on how
to make it faster.

Here is what I would like you (plural) to do:

1. Tell us when ZINC has problems. Don't wait weeks. Demand action!  Do
not wait more than 24 hours to say something is wrong. If it is wrong
for you, it is probably wrong for someone else too.  Your feedback can
help us make ZINC better.

The good news is that ZINC is better than ever, and steadily getting
better, with more compounds, higher purchasability, and, we hope, more
biologically relevant representations. Now we'll try to make it faster too.

Good docking!

John
UCSF ZINC Team


marc wrote:
> Dear Zinc Fans,
> I'm guessing and hoping this is just a transient problem, but quite a 
> few times today and over the last couple weeks I've tried copying and 
> pasting a series of ZINC codes into the "Search and Browse" utility 
> (http://zinc.docking.org/choose.shtml), pressing the QUERY DATABASE 
> button, and then getting an error saying that none of the codes could be 
> found.  I am not applying any filters, simply going to the Search and 
> Browse screen, pasting the hits, and doing the search.  The typical 
> response I get is below (bold faced by me):
>
> --- sample response---
> Performing the following mappings of retired ZINC codes... 3053242 -> 
> 6170051 2888645 -> 4821153 2798501 ->
> 4977973 5248663 -> 17837111 *No hits found* select s.sub_id,s.smiles 
> from substance s, protomer p, catalog_item ci,
> catalog c , zinctmp.qury q where s.sub_id = ci.sub_id_fk and 
> ci.cat_id_fk = c.cat_id and c.free=1 and p.sub_id_fk
> = s.sub_id and q.qury_id = 108894 and q.sub_id = s.sub_id limit 0,500
> --------------------
>
> The data I'm pasting is shown below.  I know the response is erroneous 
> because if I paste subsets of this codes (like the first couple of 
> them), the search is successful.   Furthermore, if I try the search a 
> few more times it eventually works.   Just thought I'd report this in 
> case no one else has noticed it.
> -marc
>
> ZINC08952177
> ZINC05395982
> ZINC02298507
> ZINC00867391
> ZINC05664474
> ZINC02098728
> ZINC01590394
> ZINC03053242
> ZINC12196956
> ZINC02474955
> ZINC07885472
> ZINC02888645
> ZINC06453729
> ZINC12196853
> ZINC13040948
> ZINC01080851
> ZINC02798501
> ZINC12628329
> ZINC12628329
> ZINC05028785
> ZINC08952046
> ZINC05248663
> ZINC11727470
> ZINC00869237
>
>
>
>