Print

---

OK, so the problem is due to this seemingly innocuous directive in xrootd1.xrdcfg
xrd.protocol xrootd *

What happens with this directive? It says load a protocol naming it xrootd using whatever protocol was bound to the executable (that is what the asterisk says). In this case the cmsd protocol. Since the framework automatically loads the protocol determined by the executable's name (i.e. cmsd) it is considered bound to the executable. Since the name of the protocol didn't match the one that was already loaded (i.e. cmsd); the framework loaded it once again under the name xrootd. So, internally, you actually had two cmsd protocol stacks running. Unfortunately, the cmsd protocol can only be safely configured once (it can't be run multiple times in the same process). This caused the second configuration attempt to create an inconsistent cmsd that simply would not work properly, though it wouldn't crash either.

What's the solution? Please follow these steps:

1. Kill the xrootd and cmsd daemons.
2. Remove all files in the admin path starting at /aliprod/xrd_srv/xrootd_1000/storage01
3. Remove the xrd.protocol xrootd * directive and associated comments; it is not needed in any context here.
4. Start the daemons once more.

This points out that we need to make additional checks to avoid such a problem in the future.

I will keep this ticket open until you verify this solves the problem.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or unsubscribe.

Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1