Hana 3 tier replication - logsegment growth

[former_member117942]

Hi all,

we just configured a Hana 3 tier system replication.

Database version is SPS11 (112.03)

SITEA ---------------------------------- SITE B ----------------------------- SITEC

PRD -> syncmem (logreplay) -> PRD -> async (logreplay) -> PRD

Everything works fine except for the /hana/log/PRD/mnt00001 folder on SITEC.

logsegment_000_*.dat files keep increasing (after two days there are around 500 log files) and the full disk situation is around the corner.

The same folder on SITEA and B contains only around 10 logsegment files.

Anyone able to help me in order to understand whats happening on SITEC?

Is that behaviour normal or we missed some configuration steps?

Thanks.

Regards.

[former_member205280]

Hi Maurizio,

Are you sure the log files belong to SITE C? Please check if the SITE C is a primary system.

Because based on your system landscape, the situation most likely occurs on SITE B system (it's the primary system for SITE C), especially when you're using logreplay mode without change the default parameters.

There're few parameters for system replication you might want to check:

enable_log_retention

logshipping_max_retention_size

Regards,

Michael

[former_member117942]

Hi Michael,

SAP support said us that "the issue you are running into is a known issue which the fix has not been released yet. There is a workaround to avoid running into a log full situation

Please refer to SAP note 2372278 - Log Full on Tertiary System"

And in the mentioned KB the workaround is the parameter enable_log_retention = off but in this situation is not possible to perform a fail back.

Regards.

[former_member205280]

Hey,

Thanks for the information, I didn't know that. I've encountered a similar problem before so I thought I'll just share it with you. Anyway hope they'll fix the issue very soon.

Regards,

Michael

[former_member182967]

Hi Maurizio,

Is there any alert shown in HANA studio?

What error do you see in HANA trace files (/usr/sap/<SID>/HDB<instance_number>/<hostname>/trace) of the third site.

Regards,

Ning

[former_member117942]

Hi Ning,

theres no error in the Hana Studio, all services are active and in sync.

No alert file in the trace directory, only thing i can see is a lots of messages like this in the indexserver trace...

Everything seems working correctly, except for the logsegment folder that keep growing...

Regards.