UCCE. Purge DB 80%

[CoDeC]

Hi,

we have UCCE 10.5.

we have reached 80% size of HDS database and then the purge started until the size was less of 80% (79% currently).

name	Total_Size(MB)	Data_Size(MB)	Data_Used_Size(MB)	% Data_Used
<inst>_hds	223838.63	200000	159407.19	79.7036

After the purge, we have noted that we have only the last 48h the information in TCD and RCD tables. So, I have extracted the size of the tables:

TableName	SchemaName	RowCounts	TotalSpaceGB	UsedSpaceGB	UnusedSpaceGB
t_Skill_Group_Interval	dbo	140879167	113.0761337	112.9535599	0.122573853
t_Agent_Skill_Group_Interval	dbo	48616072	30.26782227	30.26512909	0.002693176
t_Call_Type_SG_Interval	dbo	6039821	3.719299316	3.650665283	0.068634033
…	…	…	…	…	…
t_Termination_Call_Detail	dbo	234561	0.148468018	0.143615723	0.004852295
…	…	…	…	…	…
t_Route_Call_Detail	dbo	918	0.002563477	0.001678467	0.00088501

I have noted that the table SGI size is 100G, half of the database size. I have found that the retention for this table is 1095 days (3 years).

In the logs of rpl I can see that the purge due to reach 80%, the procedure purge a lot of rows of the table SGI but the size is very high. I have read in this community that the procedure of purge follow the alphabetical order.

I'd like to remove data from SGI table instead of increase the database size.

How can I purge the table SGI?  Thanks.

[Omar Deen · Accepted solution]

I implore you to re-consider increasing the size of the database... it really is the best solution.

If you cannot increase the size, the safest way to do this is to decrease the retention days. Stop the Distributor service first, go to the registery and navigate to HKEY_LOCAL_MACHINE\SOFTWARE\Cisco Systems, Inc.\ICM\<your_instance>\Distributor\RealTimeDistributor\CurrentVersion\Recovery\CurrentVersion\Purge\Retain\Interval\SkillGroup. As you already saw, the value for that key is set to 1095... go ahead and change that to something that works for you... maybe 730 for 2 years. Reboot the server and then do the same to the other side.

[CoDeC]

Thanks for answer.

I will change the value of the registry like first step. For me, doesn't make sense that the size of SGI table is the half of the size of the database.

regards,

[bill.king1]

It looks like Cisco took down the tech doc that had this process :(

https://www.cisco.com/c/en/us/support/docs/voice-unified-communications/icm-logger/20523-retention-settings.html

I didn't think a reboot was required, though?

I also thought that Cisco recommended not changing it too dramatically at once so as not to max out the Logger/HDS? So for instance to move it down 20 incremental days at a time and not 200 for instance?

[Omar Deen]

https://web.archive.org/web/20150910092905/https://www.cisco.com/c/en/us/support/docs/voice-unified-communications/icm-logger/20523-retention-settings.html

The reboot is to ensure a clean slate.

Yes, it's best to increment down, but that has more to do with how SQL replication worked and the fact that not everybody's pipeline is the same (some better than others). My experience has been fine with making dramatic changes - not during business of course.

[CoDeC]

Hi Omar,

the last night, we applied the procedure and the space released was 20%.

thanks.

regards,