US/Pisa people present at GS: Chris Orth (author of this report),
Stefano Stalio, Paulo Saggese, Danilo Mangiavacchi, Hunmoo Jeon, and
Sophia Kyriazapoulou (somewhere)
No Meeting, just reporting things as I saw them.
The shift began with many Inter-Erp overflows errors in the
BMONITOR e-mails. The problem turned out to be a miscabling on uVax3
from work done the week before, and an incorrect discriminator level on
a NIM module on uVax1. Someone once upon a time forgot to multiply
their DVM reading by 10, and we have simply been lucky not to have
errors until now.
The biggest changes this period were the addition of the
Streamer Tube Slow Monopole readout of the WFD and the addition of
STAS and QTP readout for TOHM events. So now a more general
combined Slow Monopole analysis is possible, starting with run 12300.
The WFD-Streamer readout addition went smoothly once the
author of this report reacquainted himself with the difference between
16 bit and 32 bit integers. However the TOHM readout of the QTP has
introduced a lot of NO-Q responses from the QTP. (in some runs over
2000 NO-Qs have been present throughout MACROs QTPs). Ivan Demitri (who
worked with me during this upgrade) thinks there may be some timing
issues between the scintillator and streamer tube triggers causing
this problem. The errors have an impact on the detector deadtime and
probably the data quality of the QTP for TOHM events. More work with
Streamer experts must be done to nail down these problems.
MACRO has been stopping/crashing itself almost every day these
past couple of weeks, usually with no errors. It is thought that
these crashes are related to network errors on microvax one. There
has been some speculation by A. Surdo and others that we may need to
reconfigure the network to handle a larger throughput. A DEC
service-man was called to have a look, but unfortunately instead of
arriving on Monday as planned he showed up Wednesday (see below).
Wednesday the 12th MACRO was turned off in preparation for the
installation of a new electrical breaker panel. The installation took
almost 24 hours, during which only the priveleged power to VXMACA,
VXMACB, and the lights was left on. MACRO was reawakened at about 4am
Thursday with PISA only because of unknown problems with the rest of
the detector; at 8:30 a fan unit was replaced on an Italian CAMAC
crate and the full detector came up smoothly. There seems to be some
problem with a HV channel since the power up: 5E10-1 refuses to be set
with its programmed HV value.
Other work these weeks:
The technicians have been assembling graphical maps of all of
the US electronics.
Stefano replaced the Laser Fibers to 2T10 and 1T12. There has
been no laser calibration data for these tanks in the past. Next
Wednesday will tell us if they are working.
Stop Master no Qs (mostly) during runs 12325-12340 were traced to
TWO flaky cables connecting the Stop Masters on SM5 and SM6 (both
interstop cables).
LeCroy SM2 : On Sunday morning it decided to go. When we
arrived things were okay, but later that morning our first indication
was the ERP lights all flickering more than usual. (We hadn't yet
heard from Emily). Most of the channels on the B, C, and T planes
drooped up to 200V too low. All of the reference PMTs were also too
low. We incorrectly deduced the problem was with the power supplies
and started by swapping in spare modules. The problem turned out to
be the LeCroy Controller (1445), and the run was back to normal around
2:30 that afternoon.
RISCy Stuff:
MACROs latest addition, RSGS04 has been in place in the new
lab for a few weeks now. But with the efforts of Ernesto Lamanna who
was here last week, many essential software packages have just been
installed. Ernesto has decided to set up RISCGS1, RSGS04, and RSGS02
in a server-client configuration. NIS has been installed on RSGS04,
and in the future will be installed on RISCGS1. Under this setup
users will log into both machines as under their RSGS02 accounts; the
RSGS02 macrousr is mounted on RISCGS1 and RSGS04 to accomodate this.
For continuity, users of RISCGS1 who do not have accounts on RSGS02,
or have accounts on RSGS02 with different names may want to have this
remedied.
In addition, AFS will be installed on RSGS04 (and perhaps on
RISCGS1; it already exists on RSGS02) to work with the pay service from
the caspur server in Italy. Under AFS many popular applications will
be accessible. The state of the RISC systems now can be described as
"better", but the system administrators may need some coaxing to get
NFS working properly. Files currently aren't actually available over
NFS mounts to RSGS02 - attempts to execute or copy them fail with a
time-out. Oddly enough, the NFS mount of the MACRO user disk on the
VAX systems works fine.
With the addition of RSGS04, Ernesto and others have decided
to close off access to root of MACROs RISCs. Too many problems have
resulted from well-meaning amateur sysadmins. Maintenance of the
MACRO riscs will be performed by Donati and Ernesto.