FM 5.5 OS 9.2 Slow... can I upgrade?

[juker]

We have a very large complicated database that was developed for us awhile back, it took years to fine tune but we are happy with it now. We have a network of iMac's running OS 9.2 and database access is very slow. The fewer people using it the faster it is but it's still slow. We need to fix it. What are the possibilities of getting something more modern like Filemaker 7 or 8 and running it on OS X? Will our database work with it and how do we find out? The computers in use on our network range from 5-10 at a time accessing the database all day long. I understand that OS9 may have some networking issues that make it slow?

Thanks!

Edited June 11, 2006 by Guest

[Genx]

... I wouldn't say it's the OS, it would much more likely be your network simply being clogged from lots of exchanges between server and clients... When the pipe gets to full, get a bigger one, i.e. switch to 100mbps if your not on it already or new 1000mbps .

Just check your network hardware before you go upgrading your OS's's.

And either way if you update any software it should be filemaker.

[gdurniak]

A few comments:

1. yes, check your network connections (100mbps minimum)

2. upgrade your host machine hardware, hopefully running FM Server. Get the fastest hard drive you can afford (SCSI or SATA 10K RPM)

3. turn off file sharing on the host machine

4. turn off all other apps on the host machine

5. OS 9.2 is OK, but the IMac CPUs are a little slow. Upgrading the client hardware may help (which would force you into OS X)

6. Unfortunately, upgrading to FM8 is not simple or seamless, and may require quite a bit of tweaking

7. 5 to 10 users is not that much. We have 15 concurrent users, mix of OS9, OSX, and Win XP, IMac, Mac Mini, and Dell, using FM Server 5 on Win XP. Total solution size is now over 3G

[Genx]

3 gig? What are you storing, elephants?.. i'm just curious how long its taken 15 users to get it to 3 gig...

[juker]

A few comments:

1. yes, check your network connections (100mbps minimum)

5. OS 9.2 is OK, but the IMac CPUs are a little slow. Upgrading the client hardware may help (which would force you into OS X)

This is good stuff, can you expand on a few things to help me?

In OS9.2 how do I confirm full duplex 100mbit? Secondly, we have not gotten any new computers because we were told our database wouldn't work with OSX. Is this false? Basically it would be nice to also be able to start replacing our computers with newer ones as time progresses. With 5.5 can we operate on OSX? Or maybe we can take a little step up to a slightly newer FM thjat won't require tweaking to our database but will open us up to newer hardware and operating systems? Lastly, is there any reason replacing a hub with a switch would make out network go bonkers because we tried that awhile back and things went to hell so we switched back to the hub.

[gdurniak]

1. FM5.5 works on OS X. If any problems (e.g. printer drivers) try FM6 (you can purchase FM8 with a "downgrade" license for 6)

2. A switch should work, and is more efficient (less packet collisions). If a "managed" switch, it has a web admin interface to monitor traffic

[gdurniak]

Genx,

we import various data files sent to us weekly. It took 8 years to get to 3G

[Genx]

Oh right then, imports, 8 years... fair enough.

[Genx]

That sort of really freaks me out... how do you back it up?

[Søren Dyhr]

we import various data files sent to us weekly. It took 8 years to get to 3G

This is raw undigested data, not information ...shouldn't it have been stored "as reference only"???

When you speak of slowness is it the rendering of data or is it searches and the generation of reports that is a drag to wait for.

When descriping the solution as complex is it then a plentora of fields, or is it a massive gridwork of files providing ad hoc values ?

Is summary fields and aggregation field shown live on layouts, or do they only show up when generating reports? Is the calc'fields unstored due to excessive use of globalfields and perhaps even as rule giving results based on other unstored calcfields? Are you trying to make statistical information based on data stored in repeating fields?

Why do you think the devil in the detail, must be in the hardware it's deployed on ...is the system build from a masterplan or has the process been more of an organic growth?

To the organic growth does this fit:

Novices put too much on a layout because they don’t know any better (now you do), but the reason more advanced developers do the same thing is laziness, not ignorance. Keeping layout clutter to a minimum means creating more layouts. Creating more layouts means adding to the navigation system in your solution. This is time consuming and tedious, tempting developers to just cram everything into fewer layouts. Tough! Create the additional layouts anyway.

From: http://www.foundationdbs.com/Downloads/WhitePaperForFMPNovices.pdf

...because the time to think thing's into a context often is neglected. 3-4 portals of information on a layout and aggregation of data found in these portals are going to pull out the teeth of an even mature kind of processor

...have you analysed what kind of hitting your network recieves on various layouts? ...if you havn't, then keep the layouts plain and simple to the core of the information you try to convey.

--sd