are there successful servers in the os_config run along w/ these errors?
the os_config generates an xml file from the output of the os_config script and drops it on the reports server in the data/os_config dir. my understanding is that 'new' files get put in the /new directory and then moved up one level during the etl (not 100% on that last part)
i'm not sure where the xml file is stored on the target system before it gets copied over - probably /tmp - is there anything there?
is there anything unique about the boxes it's happening on? does it always happen on the same boxes or is it random?
The problem does not affect the same servers everytime.
So I thought it was dued to a sort of "overlapping" of the threads during the parallel execution of all the os_config jobs (one per server).
So in the job configuration panel I changed the flag from 'Number of targets to process in parallel = Unlimited' to 'Number of targets to process in parallel = 20' and this fixed the problem.
By the way, it could be nice to understand why choosing unlimited this happen. We speak of just 100 servers.