10-23-2016 10:49 PM
After some major woes getting Act for web to play fair with IIS, SQL server, and backups, I have a couple of residual issues that have been here all along (three months into Act) but entirely masked by other more critical issues. These are issues revealed by perusal of the Windows Application log.
1. Periodically (once about 10 minutes after each reboot, I believe), a set of six instances of errors with source Act.Outlook.Service.Desktop within two seconds. All six show this: "Object reference not set to an instance of an object." None of the sets seem to be preceded by the same event that I can correlate as a trigger. So I cannot tell if this is important, what the cause is, or what the effect is.
2. Repeated instances of this non-error log entry: Starting up database 'ActEmailMessageStore'. I do see, once (or twice within a couple of seconds) upon reboot, the same message for each of the other Act-related databases such as model, master, ACT2016Demo, and my live DB name. But the 'ActEmailMessageStore' one just keeps on giving. Even late at night, when nobody is using Act, since a reboot half an hour ago, I have 40 instances. That seems like a lot of busy work.
On #2, I do see general SQL Express-related posts outside this forum indicating something about an AUTO_CLOSE setting; however, I certainly set no such thing, and I never knew this DB existed until I saw this string of events.
Any ideas? Does it even matter? Will this stop once users start importing e-mails into Act? (I just enabled Outlook integration last week, so I am not sure they have actually imported any yet).
08-11-2017 01:32 AM
To clarify, the email message store database handles the Act Internet Mail inbound emails on that specific machine - so on a server machine where email won't be set up it is completely unnecessary and can be detached/dropped.
10-24-2016 05:06 AM
08-07-2017 08:24 PM - edited 08-07-2017 08:24 PM
Any progress on #2? I just upgraded from v18 to v19, and the problem continues; every one to four minutes, my server logs this in the Application event log:
Starting up database 'ActEmailMessageStore'.
It is a 17137 Information message with source MSSQL$ACT7
Technically, it does no harm; however, it does very seriously clutter the event logs, filling them up so that it completely masks other items I need to find, and it still makes me wonder 1) why the system thinks it needs to start that DB every couple of minutes and 2) how one might disable the message or perhaps stabilize a process that
08-08-2017 03:02 AM
I wasn't able to find any further information on why this might be being logged so often. You could try setting the Act! email system preferences to use Internet Mail instead of Outlook temporarily to see if this makes any difference.
You can do this in Tools > Preferences > E-mail & Outlook Sync > E-mail System Setup. Click write new email afterward to open the editor.
08-08-2017 06:19 AM
This is on the server, where there should be no client-side activity and where I do not configure e-mail at all. I am trying to minimize impact Act has on the server's processor and logs.
08-08-2017 07:15 AM
08-08-2017 09:30 AM
I ran through E-mail System Setup, but nothing has made a difference.
When I check "Internet Mail", it then requires full configuration of an e-mail account, replete with specification of the SMTP server, and I have no account I want even remotely associated with a setup that is required only to stop the server from filling up the event logs. So I first tried leaving the "Internet Mail" checkbox blank, and the configuration finished without asking for more details; however, it had no effect on the "Starting up database 'ActEmailMessageStore'" filling up my event logs (and presumably tying up some degree of server resources.
So then I dummied up the information (account, SMTP, POP server, etc) to configure it as Internet Mail, but the "Starting up database 'ActEmailMessageStore'" message just keeps on repeating itself every couple of minutes in the event log.
Note that this is not a message that I see on workstations; however, on the workstations, the logs are filled with entries with Source "Desktop History Queue Provider", beginning with the time the user first opened Act after the upgrade from v18 to v19. I will post that question separately in case someone has a way to turn down the logging level for those.
08-09-2017 02:48 AM
08-09-2017 06:47 AM
Here it is. I just blanked out the User, which was the built-in domain admin account I use when logging onto this server, in this format: Domain\UserName, as well as the Computer, which was in this format: Computer.domain.
Every one of the visible log entries is exactly the same except for the date/time. I cannot figure out what triggers the event or why the gap between instances varies from about one minute to about nine minutes. Note, however, that there is a pattern here: while the number of minutes varies from 1 to 9, seemingly at random, the second slowly increases:15, 18, 19, 23, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34 36, & 41 within the screen shot. This is fully consistent when I scroll back, where I see that this loops past 00 like this: 56, 58, 59, 01, 03, 04, 06, etc. This pattern is 100% consistent as I scroll down as far back as two months (where the logs end).
That is, there is an absolutely consistent pattern here that, although the minute gap may be anywhere from 0 minutes (i.e. two instances within the same minute/second) the second moves consistently forward with each iteration. That must reveal something regarding the method by which this is being run, possibly that it is being triggered at an exact x-minute count from the completion of the last iteration rather than from the beginning of the last iteration.
But that is pure conjecture and really does not help me any. It does not appear to be a scheduled task, and the only relevant services I see are these:
Act! Scheduler: Automatic but not currently running
ACT! Service Host: Automatic & running
ACT! Smart Task Service Host: Automatic & running
I can only assume that one of the running service spawns this one.
Did I perhaps miss a note somewhere in the documentation indicating which services could/should be enable on the server side?
08-10-2017 02:53 AM - edited 08-10-2017 03:22 AM
Please could you hit Windows Key + R to open up the "Run" dialog box, then type ActDiag.exe and click Run.
Click past the disclaimer, select the Databases drop down and then select either Database Details List or Database List
You should see a list of attached Act! databases on your Act SQL Instance - could you let us know if the ActEmailMessageStore database is listed?
Edit: If the database is listed there, check to see if it is listed as recovery/suspect. It may be that the database is missing or damaged - these are all potential causes for the SQL Server constantly trying and failing to 'start up' the database.