Thank you for your post.
When a client closes a hosted file, a message is sent to FileMaker Server that the client is no longer connected to the file. If the user experiences a crash, that message is not sent, so FileMaker Server has no idea the client has left. After a couple of minutes, there is no activity from that client and FileMaker Server cannot connect to the client, so the client is then dropped.
From your post, it appears this just started occurring "several times the last 14 days". What occurred prior to 14 days ago? Is this happening for all clients or just somoe clients? Does this occur when accessing a specific file? A specific table in the file? If a specific file, run a Recover on the file and see if there are any errors in the Recover.log file.
How many files are being hosted in FileMaker Server?
How many clients are connecting to the server? Are you running web publishing?
Do you have any other applications running on the server?
How much memory do you have on the server?
Please run a Recover on a few of the active files. In addition to the Recover.log file, I would also like to get your FileMaker Server log files. Please check your Inbox at the top of this page for instructions where to send these files.
the problem is that after no activity, the server does not drop the client that have crashed. The client can login again, and I can see the client twice. The second client can I force disconnect, but not the first. Today I actually had to shut down the database to get rid of the client that originally crashed, when she opened a field with a valuelist from the contact field..
Prior to the 14 days ago, I used another server with fmps 10 and osx 10.6.8.
It have happened for two of the about 8 clients, and we are not running web publishing on this server. It is only running two databases, the datafile and the layout file.
On the server I also run Supercontainer.
The server have 4GB Ram, and is a MacMini with intel core duo 2.
I will send you the log files.
The crashing is also a concern. FileMaker files are generally much more stable than what you describe here. This may have nothing to do with the issue you describe, but you should check out your files with Recover and/or FM Diff to see if they might be corrupted. (And version upgrades, on rare occasions, have exposed latent file corruption that went unnoticed with the previous version...)
I agree with you Phil, I dont like the crashing at all. I have looked closer to the cause of the crash and maybe it was something simple as a returncharacters after the product names that is used as the valuelist field.This was found in 7 of the 1500 products I think I will have to make a scripttrigger on the field that removes return in that field. I am also consiering to avoid the valuelist and use a search scripttrigger or something else. Maybe you have some idea to use instead of the valuelist?
We had no crash yesterday, hope it helped.
But, I still dont like the ghost client in filemaker 11 that wont go away.
During the weekend I have taken down the database, done recover and found no errors.
For me it seems to happend in a script that creats a new orderline, sometimes, and more often on one user than others, but this is not the rule. The one user is also the one that creats the most orderlines.
I got this error message two times from the server, when users had been hanging, but not disconnected from the server:
FileMaker Server 184.108.40.2069 on fmpbookingservice reported the following event:
2011-08-29 00:00:40.964 +0200Error618fmpbookingserviceDatabase or temporary file "filemac:/Server HD/private/var/folders/b7/jw4rz1v52p1327bjxjnqqx9w0000gp/T/FMTEMPFM4e5ab6cf0001.tmp" is damaged and has been closed. (824)
FileMaker Server 220.127.116.119 on fmpbookingservice reported the following event:
2011-08-29 12:03:23.018 +0200Error54fmpbookingserviceFileMaker Server quitting! An unrecoverable error occurred: global temporary file damaged.
Contact information not specified.
What can this be??
I have now deleted the folder (b7) with the damaged files, and.
Any tips would be great.
I have also removed alot of calculated fields and made scripttriggers that populate previuous calculated fields with data, to increase overall speed and loadtime to optimize the system performance.
I have setup the database on another server on another location, and it runs smooth without any problems so far.
On my two servers in the office, I have had problems, one server 11 on lion and one server 10 on 10.6.7.
I am suspecting the network router or the ISP. Can it be a port conflict? It seem to crash when a certain amount of data is loaded.
We upgraded over the weekend to FM 11 across the board and we ran into this EXACT problem right off the bat. We have a crappy solution that some consultant built and I had to take over. It gets hung on some processes and the users force quit the application in task manager. It leaves their login on the server and creates another one when they login. One use had 5 client instances on the server. I even tested it on the web portal. Same thing. I was able to leave 3 logins on the server by quitting the browser before logging out. These logins stick and I can't use console to disconnect them. Nothing happens. These extra logins start to slow everything down.
We had to roll back the server to the previous version, which does not give the same problem.
Also, when they force quit (version 11), it leaves a process still running or something. When they log back in on their dashboard record and try to do anything, it says it's being modified in another window, so filemaker still thinks the last process is still running. I've experienced this on mac before in earlier versions, but not on windows.
They were running on version 9 clients before (10 server) and did not have the same problem.
We're running Windows Server 2008 with XP pro on 64 bit client machines with dual processors. The server is a virtual server on it's own box (4 processors). We just purchased a 100-user site license, so the VP of IT is annoyed to say the least.
yes the left users and multiple logins is new with fmps 11 I think. I have also experienced record locking by users that is disconnected. It can also take a long time to get the hanging users out of the server, either by closing the database, or stopping the server.
I hade a similar situation last year with users that went disconnected, but they were not left on the server (fmps 10). I hope this wil be fixed in future upgrade on fmps.
I am now happy because it is probably not any problems with my database, it runs great on the other location. I tink it is some port routing or other problems with the datastream on the router in our office, maybe in combination with the web provider.
I really would like to hear from others who have experienced crashes or disconnected users after a hang after a certain load of records or layouts. Last year a client solution caused a hang after scrolling trough abaout 18 to 20 records. And it could be reprodused. After moving to another location, it was no problem at all.
Now we had a similar situation, where the hang happened when doing certain scripts and creating a new record. After moving to another location, everything is ok.
I have two servers behind the firewall. The servers is available from outside, and I use a regular opener file.
Today I made a great discowery. We turned on ssl in the fmp server, and the database connection runs much better, no hang or beachball.
Can you understand what is the cause of this?
I also forces my fmp to quit, so my login to the server was still there after force quit of filemaker. Then I tried to relogin, but got hte message that I cant and filemaker quit as it should do. After about two minutes I also disapeared from the server.
So maybe others also should try to turn on ssl in the Database Server under security in filemaker server.
But I really would like to know what causes this.
I have the same problem but its not related to a crash. If a user disconnects ubruptly for watever reason i.e. their PC OS crashes, Network drops out whatever then FM Server still thinks they are connected. You cannot disconnect them via the admin console.
This is rinnong on Windows Server 2003
I have reported this before here Client can't be disconnected by Server if Client Crashed