BIA High Avaliability

[Former Member]

Weu2019ve noticed the fact that even if one blade isnu2019t available all the process interacting with BIA will fail, nevertheless backups seem to be configured correctly (according to information from basis team). This strange behavior appears either when initial filling of indexes or rolling up info cubes with requests.

Logs from SLG1 describe some index failures:

u201CIndex is corrupt, rollback or commit required;inde

Error Creation of the BIA index for InfoCube 'IMP_TCPPA' terminated while fillingu201D

Logs from TREX show us networks problems:

u201C[1157667136] 2008-08-19 11:27:43.784 e TrexNet EndPoint.cpp(00251) : ERROR: failed to open channel 172.31.24.149:30403! reason: connection refused

[1157667136] 2008-08-19 11:27:43.784 e NameServer TREXNameServer.cpp(00731) : pingTREXService(indexserver@sb2:30403) net exceptionu201D

Is it network or backup bug?

[former_member184494 · Accepted solution]

For us the network has been an achilles heel... also we have the same problem - even if one blade goes down we seem to lose BIA Availability. In many cases restart of the blade resolved the same.

But still BIA was unavailable for that time period.

Arun

[jgleichmann]

Hello Andrey,

what revision and backup mode you use at the moment? I have some bad experience with any backup mode in Rev47 and the bugs will only be fixed in the next revision which should be released at end of september, but i think i takes some more time.

May be this note could solve your problem 1232308.

Best Regards,

Jens