Feeder recalculation at restart with persistent feeders on

[hju]

Hi,

today i encountered a strange issue with feeders at restart of PA (11.8.00400.7-0)

With PersistentFeeders activated, the server is starting normaly loading all cubes and feeder (from file) without any issues.
There is NO ERROR message like:
TM1.Server ***** Persistent Feeder error *****
TM1.Server Feeder file error/inconsistency detected, reason: ...

After loading all cubes and feeders from file (this is the last)
TM1.Server TM1CubeImpl::LoadFeeders: Loading feeders for cube '}StatsForServer' from saved feeders file.
TM1.Cube Laden des Hauptteils für Cube }StatsForServer abgeschlossen

the server starts to newly compute all feeders
TM1.Server TM1CubeImpl::ProcessFeeders: Computing feeders for base cube ....
...

instead of executing the startup chore :
TM1.Chore Executing Startup Chore
and starting the threads before ending startup.

Problem: its a very large modell which takes hours recalculating all feeders...

Any ideas what could have caused this recalculating of the feeders properly loaded from file before without errors?

Holger

[burnstripe]

Possible reasons
A process called DeleteAllPersistentFeeders() before it was restarted
The service when restarted did not stop gracefully, aka terminated
Cube file was newer then corresponding feeder file (can happen if service not stopped gracefully)
Feeder file was missing for a cube, forcing full reevaluation (can happen if service not stopped gracefully)

[hju]

Hi burnstripe,
thx so far, but it was none of the above.

But I found a trace...
One .rux file was slightly newer than the feeder file (no idea how).
A collegue pointed out that he had a similar issue causing tm1 to recalculate "some" feeder (not all!) without logging an error.

In my case, some of the recalculated feeder (cubes) had dependencies with the cube with the mentioned .rux file.

The sever took "only" 1,5h hours to get up. He only recalculated 6 cubes.
Nevertheless its showing strange log entries:

6356 [] INFO 2024-06-20 17:06:13.672 TM1.Server TM1CubeImpl::ProcessFeeders: Computing feeders for base cube 'Artikelpreise'.
6356 [] INFO 2024-06-20 17:06:24.263 TM1.Server TM1CubeImpl::ProcessFeeders(in 10590ms): Done computing feeders for base cube 'Artikelpreise'.
6356 [] INFO 2024-06-20 17:06:33.728 TM1.Server TM1CubeImpl::ProcessFeeders: Computing feeders for base cube 'Controlling_Parameter'.
6356 [] INFO 2024-06-20 17:06:33.728 TM1.Server TM1CubeImpl::ProcessFeeders(in 0ms): Done computing feeders for base cube 'Controlling_Parameter'.
6356 [] INFO 2024-06-20 18:25:09.225 TM1.Server TM1CubeImpl::ProcessFeeders: Computing feeders for base cube 'zProcessLogs'.
6356 [] INFO 2024-06-20 18:25:09.231 TM1.Server TM1CubeImpl::ProcessFeeders(in 6ms): Done computing feeders for base cube 'zProcessLogs'.
6356 [] INFO 2024-06-20 18:25:09.238 TM1.Server TM1CubeImpl::ProcessFeeders: Computing feeders for base cube '}CubeProperties'.
6356 [] INFO 2024-06-20 18:25:09.238 TM1.Server TM1CubeImpl::ProcessFeeders(in 0ms): Done computing feeders for base cube '}CubeProperties'.
6356 [] INFO 2024-06-20 18:25:09.243 TM1.Server TM1CubeImpl::ProcessFeeders: Computing feeders for base cube '}DimensionProperties'.
6356 [] INFO 2024-06-20 18:25:09.244 TM1.Server TM1CubeImpl::ProcessFeeders(in 1ms): Done computing feeders for base cube '}DimensionProperties'.
6356 [] INFO 2024-06-20 18:25:09.259 TM1.Server TM1CubeImpl::ProcessFeeders: Computing feeders for base cube '}Process_Shedule'.
6356 [] INFO 2024-06-20 18:25:09.259 TM1.Server TM1CubeImpl::ProcessFeeders(in 0ms): Done computing feeders for base cube '}Process_Shedule'.
6356 [] INFO 2024-06-20 18:25:09.262 TM1.Transaction TM1TransactionLog::Initialize: Recovery of lingering transaction log completed successfully.
6356 [] INFO 2024-06-20 18:25:20.545 TM1.Chore Executing Startup Chore "zStartUp"

there is a big gap of 1:20 in the logs where maybe the cube with the rux file out of time is missing... its one one the big.
Still strange - but maybe i found the source of the Problem