Rhys,
> What is the recommended minimum configuration to run ACeDB (v4_9f) on
> Linux (RedHat with 2.4.18 Kernel)?
>> We have a Pentium II, 400 MHz machine with 192 MB of RAM. There are 2
> disks, with swap on one and the database on the other. The problem is
> that it takes a very long time to start up ACeDB. Once it is up and
> running, it is very quick. Pretty much all of the available RAM is used
> when ACeDB is running, but it doesn't seem that much swap is used, so
> I'm not sure that more RAM would make it start quicker. It seems to me
> that the delay is simply the time it takes to read the data from the
> disk - should we be looking at faster disk or SCSI rather than IDE?
OK, it would be useful to know some numbers here, specifically how long it does
take and also what size your database is, you can use the "Program Status" menu
button in xace to get some information here...in particular the information
about the size of the index is relevant...see next para.
As of acedb 4_9, the presence of all tags in all objects is recorded in an index
which massively speeds up queries in particular and the acedb code in general.
The downside of doing this is that index can get very large leading to quite a
long load time for acedb. This may be what you are seeing.
> Rather than buying more hardware, would it be possible, using the
> separate client and server version, to start the ACeDB server and keep
> it always running? We currently access it remotely, via X windows from
> a variety of desktops, so we don't want to just log on once and leave it
> running there. We are not concerned with allocating all of the
> resources on this machine to ACeDB, though - it is the only thing that
> this machine is used for. Generally there is only a single person using
> ACeDB at a time, and on occasion, there may be a week between use.
You could set the server running with an indefinite or very long timeout, but if
you do this you cannot currently run a graphical acedb program off it unless you
use the rather experimental sxaceclient which really can only be used for
viewing data.
I still think you may have a problem with your database though because when I
start up the C. elegans database here (which has a massive index), the wait,
while longer than one would really like, is not that long.
You can mail me direct if you like to discuss this further.
cheers Ed
--
------------------------------------------------------------------------
| Ed Griffiths, Acedb development, Informatics Group, |
| Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, |
| Hinxton, Cambridge CB10 1SA, UK |
| |
| email: edgrif at sanger.ac.uk Tel: +44-1223-494780 Fax: +44-1223-494919 |
------------------------------------------------------------------------
