[Oshino, YamaT]
Abstract
- UTP cables for the DAQ network were replaced as new ones for fixing the repeated 0xbad problem on OMC and SRs.
- k1gate was updated to a new OS (CentOS8 -> Rocky Linux8).
Details
Cable replacement
On some RTFEs, DAQ status became 0xbad in repeatedly.
This issue occurred when NIC connection was downed or link speed was changed from 1Gbps to 100Mbps.
So we had doubted UTP cables, NIC interfaces, or the network switch as the cause of the bad DAQ connection.
At the beginning of today's work, DAQ status became 0xbad on many RTFEs on C2 and C3 racks just by touching UTP cables.
There was curiously no problem on C1 racks. So we replaced all UTP cables in C2 and C3 racks.
After replacement of cables, DAQ error didn't occur even if we touch UTP cables.
I'll be watching the stability for a while.
OS update
We replaced k1gate from CentOS8 to Rocky Linux8 because the security support will expire soon.
Migration script can be available on github.
dnf repository for Rocky Linux is ttps instead of ttp.
It cannot be accessed if crypto-policies is set as FUTURE.
FUTURE seems to request more than 2048bit for certification key.
I changed the policy as DEFAULT.