corrupt native navision database

gri

hi there
we have a problem.
one of our customers, a retail store chain, has a native navision database installed in each of it's stores.
version 3.60 navision with landsteinar retail as an add-on

last week one of the stores databases crashed.
the size of the fdb file is set to 5 Gb and the average degree of usage is about 5 percent thru out the whole chain.
the corrupt database suddenly seemed to be full, 100 percent database used. this increase from a few hundreds of MB to 5 GB couldn't be linked to any unusual activity. the amount of data entered into the application was the same as in any other given day and no special jobs weren't scheduled or executed.
the database was expanded with 3 GB (huge, i know) and then the usage dropped accordingly. now it seemed that two thirds of the database size was in use ( 5/8 ). at some point during that day, the computer which serves as a server was restarted.
after restart, surprise surprise, the database used jumped to 8 Gb i. e. 100%.
i expanded it again myself with another 500 MB in order to see the miracle and i witnessed it growing again to 100 % after restart.
at this point i was sure that something in the database is corrupt and it deceives me.
after useless investigation, i've declared a disaster and we've put to work the disaster recovering policy. we've restored a back up made in the previous night and that was all.

the bad news struck us today, five days later to that incident, when we were noticed that a second database crashed in the same way.

i've searched the forums and i've found this guy which seemed to have the same issue a few years ago.

http://www.mibuso.com/forum/viewtopic.p ... se+pointer

if you could spend some minutes and shed some light into this matter, i'll be very thankfull

hope i was clear, looking forward for your replies

jm

Hi,
try to start your client with the parameter "dbreadonly=yes", then make a
backup and rebuild your database with this backup.

Otherwise you have to contact your NSC, which should contact MBS to support you.

There is a tool called C/DART to repair a database, but this is very time
expensive and will be done only by MBS.

br
Josef Metz

gri

we already did backup/restore and now it's ok from this point of view.

now the main issue is how this things happened and how can we prevent them.
the fact that two different databases crashed in the same way five day one after the other is quite frightening, because our customer has 100 databases running across the whole country.
backup/restore or any other tool for recovery would be ok if it happens once a year, once every few months maybe.
if it is happening weekly, it's mandatory to find out why as soon as possible.

and i didn't find out why, although i've checked partnersource and some other available resources.

andhilda

Hi Gri,

Do you use Multistore or some other module to "replicate" data from
POS <-> Server?? I think about Multistore. If you have this module in action then you have to look at a cleanup activity.

To me it sounds like your server/POS are sending datafiles all the time and then when the database are full the import of data stops. When you expand the database, the import process continue from where it stoped.

Look at how long time "temp sales and other data" are stored in the database and review the settings for deleting "old" data.

Best regard

Andhilda

gri

for pos and store management we're using an add-on from landsteinar retail.
data replication is controlled via Data Director, a solution provided also by landsteinar.

it doesn't seems to be caused by some sort of replication error, because even with the data director service stopped and with the database expanded, i wasn't able to insert data.

after restart the usage grows back to 100 percent, but even when the database was just expanded and it shows 60 percent usage i'm not able to insert anything.

if it has something to do with replication, i rather suspect that the replication caused some sort of inconsistency in the database at some point in time.

what do you think about that? did you encountered this sort of problems using replication of any kind.

aavio

Dear gri,

I face the same exact issue now. did you find any solution?