News: GPFS outage; GPIB; PiezoPiE871; apt clean

GPFS outage: caused a severe impact on several ExpPCs, e.g. P09 was not able to break the interlock.

 
What happened, IT (partly rephrased) 
* ~10:30 AM: Deletion of old beamtime started on beamline filesystem
* ~10:34 AM: GPFS Deadlock (between storage node and proxy) appeared for several nodes
* ~10:40 AM: First user report with beamline fs access issue
* ~11:12 AM: Deadlock vanished - without any admin interaction
--> Further investigation with the vendor

Maybe the issue is related to an automatic deletion of old beamtimes. 
Happens regularily on Wednesdays, did not cause any problems in the past.

Case will be opened at IBM, log data have been recorded.

Noticed 'Keine Stoerungen' on IT page.

Remark: to enter the experimental hutch use the Not-Auf button which unlocks the magnetic lock of the door and initiates a beam dump. So far it is not known whether the action triggered by the ICS button 'Break Door Interlock' avoiding a beam dump can be executed by some switch independent of the Web interface.

GPIB trouble: under Debian-10 it may happen that the GPIB module is not loaded properly after reboot. Tango servers communicating via GPIB will not work. The problem has been studied intensively but so far it was not possible to fix it.
User action: if one of your GPIB-related Tango servers gives problems, check 'dmesg | grep -i gpib' for 'gpib: no gpib board configured on /dev/gpib0'. If so, execute /usr/local/sbin/gpib.sh
This information is also available in the what-to-if section of the Spock manual, see here
PiezoPiE871, P21.1: reported trouble with new server, configuration problem, is running now.
apt clean: to be executed automatically not to overload /var, done.