Bahmni PAT call 24-June-2020

Reminder, we have PAT call, happening - at 6:30 pm IST, that’s 1 pm UTC, 8 am ET Wednesday, 24th June 2020.

Zoom meeting:

Please provide any topic you would like to discuss in this thread.

@ajeenckya @arjun @mksrom @pramidat @ramashish @shivarachakonda @binduak @swetha184 @laxman @anandpatel @snehabagri @sushilp @sushmit @vmalini @akhilmalhotra @dipakthapa @ramashish@pradipta @mddubey @rrameshbtech @mddubey @iadksd @angshuonline @mwelazek @michaelbontyes @buvaneswariarun @praveenad @sanjayap @florianrappl @apaule @mwelazek @som.bhattacharyya @tejakancherla @rabbott @maryp

Hi Buvaneswari,

We would like to discuss on the below talk thread.

Regards, Teja

We are facing the following situation,

Our implementation has id based sync strategy. Each location has more than 10 clinics, i. The ids and locations laid out are such that one Clinic has heavily skewed patients/encounters etc.(order of lakhs) . Other clinics are in the range of hundred or few thousand. When this clinic with heavily skewed data is being synced it always errors out. Console log shows a 503 error mostly.(Down with maintenance page) Also the sync is extremely slow.

Upon investigating we could see two kinds of issues,

i.Skewed numbers and hence many requests being sent. ii.For these servers it looks like the antivirus is eating up most of the CPU. The MTDD(Malware detection from Sophos) consumes almost 90-100% cpu when the sync is ongoing. We are suspecting one part of the problem to be that since connect on such a location is making thousands of network calls the malware detection component is intercepting all of it. We also found heartbeatd to be a CPU intensive process.

ii.Also can we uncover any cases for selective sync - based on programs etc ? We wanted to get some idea on how are other users in the community are handling it.

Please find notes here