first sync is successfull, some syncs after are ok too. from the olsync log it seems that at some point the sync is canceled (by the user or by network disc) and from the point all the handheld throws this message and cant sync successfull.
during the syncs yes data are update, sequences are used too but no autosync
I saw a note on metalink about this lock when the JVM crashes or canceled sync causes locks on C$INFO and the only way to resolve it is to reboot the device... the note was on LINUX but I recall having the same issue with Windows as well.
as i mentioned we are using handheld pda's (intermec and motorola). there is no jvm installed and used. we are using windows mobile 6.1. restarting the handheld device doesnt solve the issue. there is no session opened otherwise by restarting the handheld any opened session would be closed.
i think that i mentioned that earlier. if i remove the bin files then the data changes are not uploaded in the server. i have to manually go and do an update on the data in order for them to be resent to the server. the biggest problem is that if i do this then the next sync is a complete refresh :(.
So, now what we need to look at is why those earlier syncs are getting uploaded.
Do you have sequences that you are keeping in sync with the server? This is a problem that I had when I got the error. Also, what patches have you applied?
yes i have 5 sequences. i have window size 500, threshold 300, increment by 1 and start 1. i have not activated the checkbox for the server side sequence in any of my 5 sequences. i have applied patch 9826651 and 9767562 which includes a bug fix from another bug i had for pushing the commands from the server in the handheld.
In terms of applying the data to the server. I have seen cases where client has completed synching, but the sync is still in session on the server. The user thinks everything is ok, but about 7000 seconds later the sync finally timesout on the server and this causes all kinds of grief. Do you notice any failed syncs on the server or syncs that take a real long time on the performance tab?
this is hard for me to observe. the error occurs in our customer's system i cant reproduce it in my test environment. due to security policies i have limited access to the server unfortunatly and therefore i cannot monitor it. from the history i dont remember any failed syncs on the server at all, most times a sync fails is due to the fact that the gprs connection is not stable and therefore connectivity is low.
in the servers performace tab i do not recall any syncs that took much(expect the full refresh ones) always baring in mind the gprs connectivity provided by the cellular operator.
there is no oracle dba available there :) , i will try to schedule it someway and check it out myself. how exactly can i run the dbms_stats since i dont have much dba experience?
the users initiate the syncs mainly.
there is no automatic sync pub items
the handhelds sync either when the user presses a button or every 30 minutes from inside the application (timer based)
i have created a command on the server that executes a small secondary app on the hh. this small app will create a file. if my normal app sees this file
it will initiate a sync.
so syncs are initiated from the handheld either with or without the user's request. all syncs are done by using the syncronize function for the .net through the olite interface for .net.