Thank you for your post.
Although we have not duplicated this behavior internally, there have been a few users who have reported similar symptoms.
Your post seems to imply that the backup schedule (or perhaps any schedule) may be the culprit. Therefore, I'd like some additional information.
How often does the backup schedule run? Are you also performing a consistency check? Do the log file show the backup has completed? Do you notice the Admin Console not updating during the time of the backup, or is it at all times, and you just notice it after a backup?
For testing purposes, try logging off everyone and then logging them back on a half-hour prior to the backup schedule. After the backup schedule has completed, does the Admin Console update?
From the command line, enter "Tasklist" prior to the backup schedule. This lists all tasks that are running on Windows ("Top" on Mac). After the backup has completed, enter "Tasklist" again. Is there any difference? Are there any changes?
I know trying to send commands to the server has been a factor, so you may not be able to execute everything above.
Please keep me posted.
We are having exactly the same problem. It is very intermittent but seems to happen each week - to get the Admin Console updating again I have to restart the server which I've been doing at weekends so that I don't interrupt the users. Sometimes the problem will occur early in the week, other times it is later in the week. The problem has not occurred so far this week.
Like Constantine, the problem seems to be linked to a backup. The backup does complete and the databases continue to be served up to the clients but the admin console is stuck with the schedule 'running', the list of users becomes frozen and does not update, as does the list of databases, and sending commands such as to shut down a database fail (although using the command line works). Looking at the log, I always see an entry somewhere in the middle of the various log entries related to the backup that says:
Administrator "fmsadmin" no longer responding; connection closed
Although the admin console seems to be working fine as you navigate around it, the list of users and databases is frozen (the log viewer works) & the console is then completely useless until I restart the server.
The server has two daily schedules set up - one at 9pm which backs up all the databases with no verification and one at 3am which backs up and verifies the files. It is the 9pm backup that causes the problem. I don't know if it is related but one of the files is 2GB. Again, I don't know if it is related but looking through the server log there are lots of unexplained 'Client "xxx" no longer responding; connection closed' messages (xxx actually shows a variety of different users). These entries have either 10 or 51 as their error code.
I'll try the Tasklist command before & after the 9pm backup tonight.
I've had this happen once recently on a windows server 2008 machine. A simple daily backup that copies one small file to a shared directory (so the boss can copy it to his flash drive without shutting down the DB like he wants to) started, but never completed (according to the log.) Two additional schedules (one launches a script, the other performs a full up back up of all files on the server) then started at their scheduled times, but none were logged as having completed.
Since stopping and restarting the fms 10 server service, things have run normally.
This has now happened again, nearly a month after the previous report.
2009-09-24 17:30:00 Server Events Information Schedule "Flash Backup" running.
2009-09-24 20:30:00 Server Events Information Schedule "Delete Archived Recycler Recrds" running.
2009-09-24 20:30:00 Server Events Information Schedule "Delete Archived Recycler Recrds" has started FileMaker script "Purge If Archived".
2009-09-24 21:00:00 Server Events Information Schedule "Daily Backup" running.
Note that each scheduled task was started but none completed. Database service appears to be functioning normally, (Clients are logging on and using the database w/o trouble), but I will have to stop and restart the service directly from the Windows services control panel before these scheduled scripts will again run to completion. Note that while the log states that Purge If Archived was started. It has, in fact, not run as I can easily verify by examining the records present in the database.
Thank you for the information.
If possible, our Development and Software Quality Assurance (Testing) departments would like to see your log files, even if they are a couple of months old. I have sent you all a private message (top of this page - right side - X Messages) with instructions where to send the files. Your log filess may help isolate the problem.
I'll send you log files and a clone of the newest back up predating the incident of each file involved. The back up schedules each specify file verification, but the log doesn't verification as taking place as the "lock up" seems to precede this event. I've also had admin lock up again over the weekend, though this time the scheduled scripts continued to execute correctly even though the status for the schedule displayed as "Running".
Sometimes I can fix this type of "lock up" by simply closing and re-opening the console. In this case, no go. I then tried using the console to stop the server, but got an error message. I then was able to use the services control panel to stop the service, restart the computer and then re-open console.
I just received some additional information. If you could also provide the following:
"Can you find out details of the configurations, such as machine type, OS, total hosted file size, total RAM, total disk free space, etc."
I should have thought of the OS info myself. The email is already sent, but here's more on the system:
OS: Windows Server 2008, SP2
Software: FMS 10.0.2.206, FMP 10.2
CPU: Intel Xeon(R), ER5420 @ 2.50 GHz
RAM: 4.00 GB
64 Bit OS
C: 15.4 GB used, 4.55 GB Free = 20 GB (FMS 10 installed here)
D: 1.5 GB (Not used)
F: 78.5 GB used, 36.6 Free = 115 GB (Database files reside here in alternate folder)
FMP 10 is not normally running on this machine except for a brief time between the first and 2nd schedules when Task Manager opens a robot file to run a client side script. The file closes after script has been run so FMP runs for just a second or two and then quits.
None of these files are published to the web.
Let me know if there is any other info that will help.
I received your files. Thank you.
Your last two posts along with your files have been forwarded to the appropriate people. I will let you know if I receive any more information.
any news on this problem?
should I install filemaker on 64 bit server? or wait untill this is resloved?
Our testers have been continually working on this problem. The latest activity was from earlier this month, but this still remains unresolved.
This recent thread appears to describe the same bug. Perhaps the details reported there will supply another clue....
Constantine Pagonis wrote:
this may be a clue, in the list of schedules, backup jobs are stuck on status "running" and sending a command to run a schedule fails.
I have seen this many times, and today just happened to be another one of those situations. Indeed, there was a scheduled job (running a FileMaker script that uses server-side plugins) stuck on "running" for whatever reason. According to the logs, the job had never completed. I went to the Clients tab and disconnected the client. It disappeared from the client list, but the schedule still said "running". I then deleted the schedule and created a new one for the same script and I ran it. It again never completed, so I went to the Clients tab and there were now two "server" clients and I disconnected both of them. The schedule still said "running". I did the same thing a couple more times (I'm a slow learner).
In the end, I had no schedules and no clients in the Client list, and I'd disabled all server-side plugins...yet on the Files list, the file where the target script resides said there were four clients connected. I clicked on it to see the detail below of who the four clients were, but none were listed.
Meanwhile, looking at the data server logs via the FMS Admin Console, they were still stale and clicking the Refresh button did nothing. However, switching from 'All' to 'Warnings' and back to 'All' did update the log.
To follow up my post from yesterday, I am using FM Server 10v2 and it now looks like I'm going to have to restart the server to gain any type of control over this thing. The database and web server part of it is all working fine. But the Admin Console and fmsadmin command line are both dead. No database access log or other database log entries have been made since yesterday (checked via the Windows event logs, as well; web logs all appear fine). Disabled and re-enabled logging -- still nothing. Last night's backup schedule still listed as "running". Tried the command line 'fmsadmin list clients', it asks for username and password, but then it freezes before it gives me the list. Just tried the same to list schedules, but with the same results.