I just saw this in Notices tab of BOINC Manager. What is it?

   BOINCstatsBAM!: Notice from BOINC

   Message from account manager: StartTag: invalid element name

   . 

   Premature end of data in tag url line 235

   . 

   Premature end of data in tag project line 234

   . 

   Premature end of data in tag acct_mgr_request line 2

   .

   4/19/2015 7:59:25 AM

Right now it is 8:09am for me, so this happened during my last hourly automatic Account Manager update about 10min ago.

Another close by machine I checked doesn't show this even after I forced a sync with BAM!.

Probably exotic characters in the hostname, username or teamname or a task with such characters.

You can use an online XML validator to find the exact problem.

Probably not. I have no funny host/user/team/project names, and the section of the XML file (acct_mgr_request.xml) doesn't seem to have any task names in it. Of course, since this report, another BAM! update happened with no error, so the XML I have now is without the problem. Looking at the lines around 234 in the new file, I see:

   <project>

      <url>http://lhcathomeclassic.cern.ch/sixtrack/</url>

      <project_name>LHC@home 1.0</project_name>

      <suspended_via_gui>0</suspended_via_gui>

      <account_key>blah blah blah</account_key>

      <hostid>10356004</hostid>

      <not_started_dur>0.000000</not_started_dur>

      <in_progress_dur>0.000000</in_progress_dur>

      <attached_via_acct_mgr>1</attached_via_acct_mgr>

      <dont_request_more_work>0</dont_request_more_work>

      <detach_when_done>0</detach_when_done>

      <ended>0</ended>

      <resource_share>100.000000</resource_share>     // This is line 234 right now

   </project>

So, at best, BAM! received a truncated file (network hiccup between Seattle and Netherlands?). At worst, BAM!'s buffer while parsing the XML got corrupted.

And, I looked at a few of my other machines. Around half seem to have a similar error report under the Notices tab, at random dates, in random places of the same file. Given that I don't have funny host/user/team/project names and given that this file doesn't carry any task names, and given that our home gets really high quality 175Mbps connection with no data loss I can remember, ever, and given that they are all running the same version of BOINC churning data for the same projects, either BOINC Manager is sometimes screwing up creating and/or transmitting that file to you, or something bad is happening randomly on the BAM! side (file is not getting to you correctly through the datacenter pipes, some buffer overflow happens, some pointer dangles, some other error falls through and gets reported wrongly as an XML parsing error, etc). So, this might actually be an indication of a larger problem.

May I humbly suggest that when you are about to return such an error back to the BOINC Manager, you also create a backup copy of the incoming acct_mgr_request.xml that you didn't like along with the error message you returned, stamped by the date/time (assuming you are not already doing this)? If space is an issue, just limit it to the last N of such errors. This way you can take a quick look at the file and determine if they really are bad coming to you, or the handling of it is broken.

Just a thought...

Tuna

This indicates that global_prefs must be kept in the Einstain@home.

Respectfully: Huh?

That's seriosly. Before connecting to BAM, you have move global_prefs to right location.
For instance, WCG and SAT@home returns invalid XML structure after manipulating "create/remove additional preferencies for school, for work..".
BAM doesn't validate all XML fragments and tags. So, you have broken BAM XML request/reply.

Using Boinc Manager, you have ability to view number of invalid/uncknown xml-tags in scheduler replay for particular projects.
Boinc Manager -> BOINC diagnostic Log Flags: switch ON "unparsed xml". See number of invalid tags at Boinc Manager->Event viewer.

* i'm always have global_prefs on Einstain@home