timeos

I cant believe that such small photos (up to few KBs) can cause such huge memory consumption. Even if we are talking about 120 users totally. Yes, I can allocate...

ok, I will do test this evening - I will allocate 5GB of RAM to this system and will monitor tomorrow how it will behave.

ok, seems this is definitely some memory leak - today I have assigned 5GB RAM in total, enabled jpegPhoto LDAP VCARD attributes and start monitoring. At beginning I was able...

yes, it could be correct allocation pattern, but this huge allocation happened just because I fetch info about one single user - I was seeing this memory allocation progress while...

while I wrote previous message, the memory consumption started again to increase with same pattern up to 3GB just for beam.smp and then it was freed again to ~900MB. seems...

another-related issue: for example today, I am unable to fetch user info ("Information: Retrieving...") and one CPU Core is at 100% all the time. beam.smp is all the time running....

not sure if it can help anyhow, but I have just did strace for the most utilized process of beam.smp. attaching here 30secs long strace. [utilized-beam.smp.zip](https://github.com/processone/ejabberd/files/259136/utilized-beam.smp.zip) Strace for beam.smp process...

Now I get feedback from users, that some of their messages to another users within same domain are lost - they did not simply arrived to recipients. this again started...

Ok, one more report from me: since I disabled photos last afternoon in mod_vcard_ldap, system is constantly idle with load: load average: 0.00, 0.01, 0.05 and with constant memory consumption...

@Rado: seems this is really the case. In our DB we had two contacts with avatars in BMP format instead of JPEG. Now I have converted them, enabled back photos...