Re: Sql ce server process

Brandon,

It seems I misunderstood what you were doing. You do have a lot going on, but don't let me discourage you from trying some new tricks. To answer your original question about is anybody else doing this, maybe someone will reply. I just don't recall seeing any other messages attempting to do what you're considering. Please let us all know if you find a soltuion that you're happy with.

--
Ginny Caughey
Device Application Development MVP

www.wasteworks.com
Scalehouse and Billing Software for Waste Management



"Brandon" <Brandon@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message news:02F0F2A8-178F-4154-9819-69FF372F5BA2@xxxxxxxxxxxxxxxx
Thanks again for the post -

We are certainly aware of the dll load issue, and have strategies in place
to ensure that this is not a problem. We have a startup process that uses
LoadLibrary/FreeLibrary to strategically order all of the dll's that will be
used by the applications configured on that device to minimize fragmentation,
and then each individual process uses LoadLibrary to maintain a static
reference to all of the dll's that it will use in order to prevent
VirtualAlloc from allocating heap/stack in the space that the needed dll's
map to, causing the dll to fail to load when it is needed. When we have
problems we are actually catching a SqlException that says "not enough
storage" (can get more details if interested) when we execute the command (I
have found many of the error messages to be misleading when you are running
up against memory limitations...I used dumpmem to look at the VM space of the
process as this was occurring, and saw that there were no contiguous blocks
in VM that were large enough for the minimum allocation size (188K)).

I'm not sure exactly what you are getting at with the 2 processes...we
actually have about 6 long-running processes - two native processes that
comprise the turn-by-turn navigation software that uses the speech engine, a
.net process that maintains network connectivit/least cost routing, a .net
process that locks the device down and only gives the user access to what
they need acces to, our .net replication process, a .net process that
handles serial communications with GPS devices, and our .net end-user
application that is using sqlmobile and is on the virge of disaster. Most of
these process seem to have plenty of free VM, just not the end-user
application that uses sqlmobile.

"Ginny Caughey [MVP]" wrote:

Brandon,

It sounds like you've done a lot of thinking about this. I definitely
understand about needing to support those WM2003 devices. Have you tried
running 2 processes at the same time, one that uses the SQL Mobile engine
and one that doesn't to see if that's a solution? I'm skeptical, but then I
haven't actually done the tests. The main thing I'm aware of with the
unmanaged SQL Compact DLLs is that you need to get them loaded early in the
app's execution to make sure there is space for them. The managed code parts
can swap in and out, but if there's not enough memory left when you attempt
to load the engine, then you do have a problem.

--
Ginny Caughey
Device Application Development MVP

www.wasteworks.com
Scalehouse and Billing Software for Waste Management



"Brandon" <Brandon@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:F10F5CE3-CA03-4E4C-AFC8-62D65DAF9C2B@xxxxxxxxxxxxxxxx
> Target platform is currently WM2003, and it is understood that future
> platforms may have sql in ROM - which you are correct would make this
> solution a moot point. We are unlikely to target another platform > until
> WM7
> is out, which takes the whole VM and process count limitation issues > out
> of
> the picture, but the devices we are deploying on today are expected to > be
> around for the next 5 years or so, so we have to consider ways to > continue
> to
> add functionality in this constrained environment.
>
> As for the reliable connection - the engine, running on a particular
> device,
> would only be interfaced by a process (or processes) running local on > that
> device - we would not in general use this to interact with the database
> remotely, though this may end up being handy in certain support
> situations.
> My plan is not to try to use sql in any more robust of a fassion than > was
> designed - only to extract that functionality into a separate process
> where
> it has more freedom to roam in it's own VM slot. I do not expect to > put a
> load on the database that is any more heavy than the load that exists
> today,
> when the application uses the sqlmobile API in-process (actually > probably
> less load, since the new process should have some free VM to cache > query
> plans).
>
> We are definitely focused on the occasionally connected end of things -
> our
> devices are in vehicles roaming the country on GPRS... This is one of > the
> reasons we exported all of the replication logic into it's own process > in
> order to handle a lot of the retry logic without bogging down the > user's
> interface.
>
> "Ginny Caughey [MVP]" wrote:
>
>> Brandon,
>>
>> It's not necessarily a bad idea, but I'm not sure if it's a great one
>> either, and it might be more complicated to implement than you expect.
>> The
>> main problem I see with that approach is that it requires you to have >> a
>> reliable connection all the time or you're completely out of business. >> I
>> don't know what type of hardware you're targeting, but SQL Compact is >> in
>> ROM
>> beginning with Windows Mobile 6, so the other 'problem' might be that
>> newer
>> devices solve this problem for you, or at least this solution might >> not
>> actually solve your problem And a third issue might be serializing
>> access
>> to data if more than one device might be using this service. Making >> that
>> robust could be more trouble than it's worth, and making it performant
>> under
>> load might be impractical.
>>
>> As for why somebody hasn't done this already - Microsoft has been
>> focusing
>> more on the occasionally connected scenario with Sync Services for
>> ADO.Net.
>> The smart device piece of that isn't available yet, but Sync Framework >> is
>> certainly the direction they're moving in. Steve Lasker even has a >> blog
>> post
>> describing a "stored procedure" kind of approach to SQL Compact:
>> http://blogs.msdn.com/stevelasker/archive/2008/02/11/stored-procedures-and-sql-server-compact-the-great-debate.aspx
>> Definitely worth a read since it addresses working with queries in a
>> uniform
>> way, although I don't think it addresses the specific issue you're >> trying
>> to
>> solve.
>>
>> -- >> Ginny Caughey
>> Device Application Development MVP
>>
>> www.wasteworks.com
>> Scalehouse and Billing Software for Waste Management
>>
>>
>>
>> "Brandon" <Brandon@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
>> news:09E3294C-67CD-4454-8F15-3C69A9459806@xxxxxxxxxxxxxxxx
>> >I posted this on the compact framework board, as the reasoning behind
>> >the
>> > direction described here is compact framework - related, but I >> > though
>> > there
>> > might be an expert over here on this board that might be more >> > familiar
>> > with
>> > possible past attempts in this direction....
>> >
>> > We are running out of VM in our main application, and among other
>> > things,
>> > I
>> > am considering how we might break functionalities across groups of
>> > native
>> > dll's that have to load for those functionalities. An obvious break
>> > line
>> > involves sqlmobile (3.x). I threw together a quick proof of concept
>> > for
>> > basically creating a database engine process that would perform all
>> > sqlmobile
>> > access, with a TCP interface for issuing commands and getting >> > results
>> > wrapped
>> > in XML (we already do something similar in a process that performs >> > all
>> > of
>> > our
>> > replication - conceivably, this process would grow up to be a >> > sqlmobile
>> > engine that handles both queries and replication). I know that >> > there
>> > may
>> > be
>> > something we can do that will be more efficient than XML, but our
>> > result
>> > sets
>> > are generally small - probably the biggest one is 18 rows by about >> > 30
>> > columns. Hopefully an added benefit would be that we might even be
>> > able
>> > to
>> > put together an algorithm for caching a few query plans...we have >> > been
>> > unable
>> > to do this for some time, due to VM constraints.
>> >
>> > I can't imagine that this is as avenue that has not been explored, >> > and
>> > am
>> > just wondering if anyone has any comments on the approach - is this >> > a
>> > bad
>> > idea? It seems like if this were a viable option, that it would
>> > already
>> > exist out there somewhere, probably even provided by Microsoft. >> > Maybe
>> > they
>> > were just afraid people would try to use it in a more robust way >> > than
>> > it
>> > was
>> > intended...?
>> >
>>


.

Relevant Pages

  • Re: Sql ce server process
    ... The main thing I'm aware of with the unmanaged SQL Compact DLLs is that you need to get them loaded early in the app's execution to make sure there is space for them. ... The managed code parts can swap in and out, but if there's not enough memory left when you attempt to load the engine, then you do have a problem. ... > involves sqlmobile. ...
    (microsoft.public.sqlserver.ce)
  • Re: Sql ce server process
    ... We are certainly aware of the dll load issue, ... application that is using sqlmobile and is on the virge of disaster. ... running 2 processes at the same time, one that uses the SQL Mobile engine ... unmanaged SQL Compact DLLs is that you need to get them loaded early in the ...
    (microsoft.public.sqlserver.ce)
  • Re: Diesel electric
    ... Your second engine will then be down as well. ... A big generator is going to be noiser than a small one. ... Unless your hotel load ... running with diesel eletric propulsion. ...
    (rec.boats.cruising)
  • Re: Diesel electric
    ... Spend my money on a good main engine and a nice small generator and saving a lot for extended cruising time;-) ... If a diesel used in auxilliar sailboat service doesn't run, it is as apt to be a fuel problem as anything else. ... Unless your hotel load approaches 50% of your propulsion load running just the one generator for ship's service will probably result in it's running too lightly. ...
    (rec.boats.cruising)
  • Re: Carburetor Theory: Idle Mixture
    ... >>> Why does a load on an engine with a lean mixture cause the engine to ... >> to develop enough power to take up the load, ... > drop, to say 650 rpm. ... > sophisticated idle jetting arrangement in the carb. ...
    (rec.autos.tech)
Loading