Software

What where

On archive

  • Debian 12

  • postgresql 15

    • main DB, synchronised to Potsdam

  • npreal ttyS

On wifsip

  • Debian 12

  • GRB Socket server receiving triggers from GCN

On sky

  • Debian 12

  • docker containers of mqtt, telegraf, influxdb, and grafana

  • matterlog logs to mattermost

  • lightmeter SQM software

On stella.vm

  • debian 12/13

  • postgresql 17

On stella-grafana.vm

Others

papaya is still running Debian 8.11

Things to watch out for

rebuild moxa drivers

A very common error is that the kernel has changed and no moxa driver modules have been built yet.

Hints are errors in the log files of problems accessing ttyS devices.

On each of the 3 computers (archive, wifsip, sky) there under /root/moxa is the current driver set, which needs to be rebuilt by running ./mxinst. Options to look out for is [n] for ssl.

restarting /etc/init.d/npreals will fix things, but all services need to be restarted.

SMART and RAID status monitoring

Currently used are 8-port megacli RAID cards: Reference: https://wikitech.wikimedia.org/wiki/MegaCli

Example commands are

* megacli -CfgDsply -aALL
* megacli -L0 -a0
* megacli -L1 -a0
* megacli -CfgDsply -aALL | less
* megacli -EncInfo -aALL
* megacli -PhysDrv0:0 -aALL
* megacli -PhysDrv0:1 -aALL
* megacli -PhysDrv0:0 -a0
* megacli -PhysDrv0:1 -a0
* megacli -PhysDrv 0:1 -a0
* megacli -PhysDrv [0:1] -a0
* megacli -PDList -aall | less
* megacli -AdpAllInfo -aALL | less
* megacli -AdpSetProp AlarmSilence -a0
* megacli -AdpAllInfo -aALL | less
* megacli -CfgForeign -Scan -aALL
* megacli -pdInfo -aALL
* megacli -pdInfo -a0
* megacli -pdInfo -PhysDrv [252:0]  -a0
* megacli -pdLocate -PhysDrv [252:3]  -a0
* megacli -pdLocate -start -PhysDrv [252:3]  -a0
* megacli -PdGetMissing -a0
* megacli -AdpSetProp AlarmSilence -a0
* megacli -PDList -aall | less

To replace a drive, do the following:

* megacli -PDOffline -PhysDrv [252:6] -a0 (if not already dropped out)
* megacli -PDMarkMissing -PhysDrv [252:6] -a0 (optional, did not work for me)
* megacli -PdPrpRmv -PhysDrv [252:6] -a0
* replace drive
* megacli -PDHSP -Set -PhysDrv [252:6] -aN (set new drive as hot spare)

To check SMART status

* megacli -PDList -a0 | grep "Device Id"  (find ID for the drive in question)
* smartctl -a -d megaraid,12 /dev/sda

Firewall

Bering uClibc 7.x firewall running also the wireguard VPN server.

https://bering-uclibc.zetam.org/wiki/Bering-uClibc_7.x_-_User_Guide

Workflow

  • ssh fire

  • edit /etc/shorewall/rules|policy or wireguard settings

  • exit lrcfg and test the rules with /etc/init.d/xxx restart

  • if no errors, can start lrcfg again and _s_ave settings, maybe _c_heck first what is being saved

  • can also run upgrade -c -vvv to check if upgrades are available

Interfaces

  • eth0 is uplink, fixed IP, ip_proxy_arp=1

  • eth1 is loc, public IPs, dhcp, no natting

  • eth2 is lab, dhcp, NATted, and currently offline