CERN Accelerating science

HasRedirect ,

Hello,
Today new entries appear in the MGM log with “func=HasRedirect”. This is something new (at least it did not occur in the coarse of the whole month of November). This corresponds to a test of the ALICE VO (ADD test) and the test fails with “[FATAL] Redirect limit has been reached” … This is since 4:00AM today.
Does someone know in which cases this “func=HasRedirect” is used ?
Thanks
JM

Hi JM,

This means you have a redirection rule in the access list. Just do “eos access ls” and it should be clear where this comes from.

Cheers,
Elvin

Hi Elvin,
On the master (currently naneosmgr02):

[root@naneosmgr02(EOSMASTER) ~]#eos access ls

On the slave:

[root@naneosmgr01(EOSSLAVE) ~]#eos access ls

Redirection Rules …

[ 01 ] ENOENT:* => naneosmgr02.in2p3.fr
[ 02 ] w:* => naneosmgr02.in2p3.fr

But I think these are normal…

The problem went away after I performed a MGM role exchange. I suspect some kind of communication problem inside the cluster but I have not been able to find a clue in the logs.

JM

Hi JM,

Yes, these rules are expected. If you see this again run the access command and post the output. It could be that the MGM you were looking at was a slave or some switch happened during that time.

Cheers,
Elvin

Hi all,
I stumbled upon this by accident, but I have a question about these access rules. Is this basically the redirector subscription to the regional/global redirector (for CMS)? I think we are missing these. Is there. more information on which of these to set and also to which redirector?

(sorry for hijacking this)
Best,
Erich

Edit: I’ve come to understand that this is also part of the internal redirection mechanism of EOS, I’ll open an extra topic for this, Subscribe to global / regional redirector (CMS)

Hi Erich,

This is not connected to the redirector thing you mentioned, I will reply to your question in the threads you opened.

Cheers,
Elvin