CR1 Detector FLPs

OBSOLETE: This page is now obsolete, new location at https://alice-flp.docs.cern.ch/Advanced/P2DetectorFLPs/

General information

The CR1 detector FLPs are the final setup, currently being used to perform detector commissioning activities for the following detectors:

DetectorFLPsCRUsC-RORCsCPUHostnamesHead nodeAnsible InventoryFLP suiteYUM update
CPV11SILVERalio2-cr1-flp162alio2-cr1-flp162inventory-p2-cpv0.17.027/1/2021
CTP11SILVERalio2-cr1-flp163N/A
DCS1SILVERalio2-cr1-flp199By reinstallation 16/2/2021
EMC28SILVERalio2-cr1-flp146..147alio2-cr1-flp146inventory-p2-emc0.16.0
FIT13SILVERalio2-cr1-flp180inventory-p2-fit0.17.01/2/2021
HMP24SILVERalio2-cr1-flp160..161alio2-cr1-flp160inventory-p2-hmp0.17.022/1/2021
IB driver installed 25/2/2021
ITS1224GOLDalio2-cr1-flp187..198alio2-cr1-flp187inventory-p2-its0.17.322/1/2021, IB driver updated 27/1/2021
MCH1130SILVERalio2-cr1-flp148..158alio2-cr1-flp157inventory-p2-mch0.16.0
MFT511GOLDalio2-cr1-flp182..186alio2-cr1-flp182inventory-p2-mft0.17.227/1/2021
MID12SILVERalio2-cr1-flp159alio2-cr1-flp159inventory-p2-mid

0.17.0

22/1/2021, IB driver installed 29/3/2021
MVS4?3SILVERalio2-cr1-mvs01..04alio2-cr1-mvs01inventory-vertical-slice0.17.2
PHS24SILVERalio2-cr1-flp164..165alio2-cr1-flp164inventory-p2-phs0.17.027/1/2021
TOF24GOLDalio2-cr1-flp178..179alio2-cr1-flp178inventory-p2-tof0.17.0
TPC145361SILVERalio2-cr1-flp001..145
(alio2-cr1-flp145 special)

alio2-cr1-hv-head01

inventory-p2-tpc0.17.027/1/2021
TRD1236SILVERalio2-cr1-flp166..177alio2-cr1-flp166inventory-p2-trd0.16.01/2/2021
ZDC11SILVERalio2-cr1-flp181alio2-cr1-flp181inventory-p2-zdc0.17.026/1/2021
Total19947416

All hosts which are currently in use for commissioning (TOF, MCH, EMCal etc…) have also been aliased to aliflpcom-XXXN where XXX is the name of the detector and N is a sequential number (only for detectors with multiple FLPs).

All hosts have an IPMI interface reachable by adding “-ipmi” to the hostname (e.g. alio2-cr1-flp001 => alio2-cr1-flp001-ipmi). Username is root and the password is NOT the same as for the equivalent Linux account (ask the CR1 support team). The same “-ipmi” host can be reached, same credentials, via HTTPS for an interface which is much more user-friendly compared to the SSH server and has real-time console-sharing capabilities (it is basically the same thing as being in front of the physical machine, very handy to see what is going on).

CR1 currently runs behind an HTTP proxy. The address is the following (both for HTTP and HTTPS):


10.161.69.44:8080

O2/FLP ONLY: in Foreman (https://alio2-cr1-prov01-mgmt.cern.ch/) all the detectors are defined as search bookmarks (including the vertical slice) in the “Hosts”/”All hosts” page (login using your CERN credentials).

For operations not using the standard FLP setup procedure, the environment can be configured manually via the commands:


export HTTP_PROXY="http://10.161.69.44:8080"
export HTTPS_PROXY="${HTTP_PROXY}"
export http_proxy="${HTTP_PROXY}"
export https_proxy="${HTTP_PROXY}"

The machines are visible from GPN. If you see any issue with network connectivity (BOOTP timeouts, DHCP timeouts, no connection, slow internal connections) contact the flp support team (alice-o2-flp-support@cern.ch).

Re-installation of CR1 detector FLPs

Ansible inventories

The examples use the existing per-detector Ansible inventories (defined here), which is the preferred way of installing CR1 machines because they define several important variables. The inventories names follow the following syntax:


inventory-p2-detectorName

where detectorName should be replaced by the 3-letter name of the detector.

If some nodes need to be excluded from the inventory, the –limit option can be used as described here. For example:


... --limit "all:!excluded-1:!excluded-2"

1. Select nodes for re-installation in the CR1 Foreman instance

Select the nodes to reinstall and change them into build mode.

2. Install o2-flp-setup


bash <(curl -s https://cernbox.cern.ch/index.php/s/kNMAmYaN9l6RaKD/download)

More detailed documentation here

3. Define the inventory

Either in the shell setup (e.g. in ~/.bash_profile) or in the running shell (in which case you will have to repeat the definition for each new shell e.g. when you logout or reboot the host you are using) define the detector and the directory either giving the ansible inventory or explicitly.

Via ansible inventory:


declare -x INVENTORY="-i ${HOME}/.local/share/o2-flp-setup/system-configuration/ansible/inventory-p2-detectorName"

Replace “detectorName” with the 3-letter name of the detector you are working with. See above for optional exclusion of selected nodes.

Explicitly:


declare -x INVENTORY="--head HEADNODE --flps FLP1,FLP2,FLP3,..FLPn"

4. Setup the HTTP proxy

If you run the installer from a dedicated controller node:


declare -x O2_FLP_SETUP_PROXY="10.161.69.44:8080"

If you run the installer from a target node, you should add the following option to the deploy commands on steps 7, 9,11 and 12:


--proxy 10.161.69.44:8080

5. Reboot the nodes for OS re-installation


o2-flp-setup reboot ${INVENTORY}

NOTE: the script will wait for the machines to come back up, you can avoid this giving ^C.

NOTE: if you are running the installer on the single target node, it may be easier to use the "reboot" command.

6. Select the FLP Suite version to install

If you want to use a version of the FLP suite different from the latest tagged one (already fetched at the end of step 2), download it out via the checkout command specifying the tag name e.g.:


o2-flp-setup checkout flp-suite-v0.11.2

or


o2-flp-setup checkout dev

7. Execute post-installation actions (e.g. create user accounts and configure the IB card)


o2-flp-setup deploy --modules post-installation ${INVENTORY}

8. Reboot machines


o2-flp-setup reboot ${INVENTORY}

or directly using the reboot command for single-node setups.

9. Deploy FLP Suite


o2-flp-setup deploy ${INVENTORY}

10. Execute detector-specific actions


o2-flp-setup deploy-detector --detector detectorName ${INVENTORY}

NOTE: replace detectorName with the 3-letter name of the detector you are working with.

11. Reboot machines


o2-flp-setup reboot ${INVENTORY}

or directly using the reboot command for single-node setups.

12. Check the installation


bash -c 'declare -x O2_FLP_SETUP_PROXY=""; o2-flp-setup diagnose ${INVENTORY}'

13. Optional: deploy the flp-dev role


o2-flp-setup deploy --modules flp-dev ${INVENTORY}

Re-installations

During re-installations it is possible to skip the removal of already installed packages using the option:


--skip-modules=o2-cleanup

Admin operations

The Ansible inventory for the vertical slice allows to perform administrator operations on all or part of the setup.
It defines the first node as head and all nodes (including the first one) as FLPs.

Prerequisites

1. Install o2-flp-setup (instructions here)

Examples

Check connectivity to all nodes


o2-flp-setup ping ${INVENTORY}

Reboot nodes


o2-flp-setup reboot ${INVENTORY}

Update all packages on all nodes


o2-flp-setup sysupdate ${INVENTORY}

Useful links

o2-flp-setup installer

Optical Fiber Infrastructure

New optical fiber infrastructure between detector frontend electronics and FLP in CR1:

– Trunk cables and subracks infrastructure between cavern/CR4 and CR1 (see attachment).

– Patch cords mapping in the cavern and CR4.

– Patch cords mapping in CR1.

CR1 Layout

Schema of CR1 layout with FLPs and CRUs/CRORCs distribution and IPMI/fiber/InfiniBand network.

LTUs config

First column: detector, second column: LTU number.


fv0 37 c25t 105 106 
ft0 38 c25t 113 114 
fdd 39 c25t 115 116 
its 40 c24t 101 102 itsmft 
tof 41 c24t 103 104 
mft 43 c24t 105 106 itsmft 
mid 44 c24t 113 114 
trd 45 c26t 101 102 lvpecl 
hmp 46 c26t 103 104 lvpecl 
phs 47 c26t 105 106 lvpecl 
emc 48 c26t 113 114 lvpecl 
cpv 49 c26t 115 116 
mch 50 c24t 115 116 
tpc 51 c24b 101 102 
zdc 52 c24b 103 104 
ticg 42 c25t 103 104 
ctp 64 c25t 101 102 
gbtit 53 c24b 
ttcit 8 c26b 101 102 
#oldttcit - c??? vmeaddr: 
# PDU: alidcsaux021, alidcsvme017 ->ttcitsw.py 
spare18 - c24b 
daq 23 cr1