Search results for query: *

7.4 and all of a set of views that were created (mostly interdependant) cannot be recreated - they must be done by hand after a restore.

That's the other problem with postgres - the pg_dump app is screwed when it comes to views - back them up separately (I am working on a solution at the moment - if it works, I'll give you the shout). The problem being that if a view is dependant on another view, the dump app ignores this, but...

has the link table got the correct details for the field, notices pk, gets correct data type,etc?

Right, here we go - 1. Error messages in unpredictable places 2. Yes, and No. Some are, some are more generic. They look like they are ODBC attempts to ttry to connect to the serbver, and failing. 3. yes, they're exactly the same as royceybaby above (pasted here to safve scrolling: 4. Fails...

I have taken a copy of the nightly backup and dumped it onto out stage environment (close to live hardware-wise) and tested it with no errors.

Right - my perfect user has problems - it's probably not a random hit, rather som users haven't hit it yet. We can't shift our mysql off the server, but it's not very intensive, and does not communicate at all in the same place. The mysql gathers site stats, and talks to the web only, whereas...

sadly not - it's the live webserver!!

I've just had a thought - gwinn7's idea (suggestion about the nic/cabling) was great, but saldy not the solution. There is another db (mysql) running on that server and it has not had any problems as of yet. Cheers anyway Gary.

I will get a light user to start using it for long periods of time now. Cheers

The machines are of different Win versions (no pattern), different Office versions (no pattern), but otherwise have identical 'other' software on. They all have same activex versions, odbc, and there does not appear to be any correlation there. There are also different levels of use for each of...

well, it doesn't seem reasonable, no - but I can't say I've had those issues with MySQL before. Apparently there are plans to make a strict casting switch - I assume that this would stop the problem, as it wouldn't be able to cast the value.

This is not an issue at all, if no unexperienced users are allowed access to the backend (unfortunately we have one in our company )

No problems - although there is an itching feeling in my head that there are a few like that one, but they're not coming to mind. If i recall them, I'll let you know. WonderCow

Sorry about this, but there is so much info I have gathered over th last two weeks, I haven't given all the details - and one thing I said was a little misleading: The machines affected were always the same ones, but they were random in the sense that there was no pattern to work out why some...

mem is 4497000k used : 138000k free disk usage is 23% used (17G total disk size) I'm unsure about the amount of mem - think it looks ok though, but the disk space seems Way within bounds.

Actually, i've not checked disk/mem usage - I will do so now...

Hey dude, Although Postgres is a good solution - you must be careful with anyone who has access to the backend - If you try to set a primary key to nextval('schemaname.sequence'::text) (autoincrement...) users can insert whatever value they wish, I have not been able to keep it strict. If the...

There are no major server processes - we are running 4cpus which are averaging: 1 99.4% idle 2 84.4% idle 3 95.0% idle 4 93.2% idle So far, I have tried the following: Checked ODBC timeout - OK Check Driver version for ODBC Postgres - OK Check ODBC Settings - OK Check for DB Corruption -...

I have set postgres loggin for the client machines, but I am now convinced that it isn't thwe code, since someone else is getting these problems, and he's not using the app. if it is client side, I am leaning more toward odbc problems, although I don't really think that it is. I have been...

Thanks for responding, but the problem is a little more odd than it looks: We have been using this combination for over a year at least, and have not had this problem. The problem started a month ago, and we have changed nothing that we can think of. Sadly, the users didn't bother to tell us...