Ispcofig on Azure, Dynamic Memory.

I've got an instance of Ubuntu Server 18.04, got it running on with the perfect server setup. Set up a mirror server in another datacentre and got it running as well as per the multiserver tutorial, except that I'm using Galera and a third database server to run an SQL cluster. It's running great. Really fast and I'm supper happy with it. Bar one little issue at the minute on the Azure instance. I'd like to use the dynamic memory function. So that it can grow the memory underload if it needs it. Starting at 2GB and growing to 4GB or more if need be. I've got other servers running with dynamic memory and the grow and shrink all the time with no problems, except my Ispconfig instance. It just grows until it uses up all the memory. Then it starts sending me warnings that it is hungry for more. I've done all I can to narrow it down using my limited Linux knowledge and I'm sure it's all being used by something filling the cache. But I don't think it is the regular disk cache that's filling it, seeing that disk cache is enabled on my other Linux servers and there not filling up[in saying that, they have smaller data sets so I can't be sure]. I know that this is out of the scope of this forum and you'd be in your right not to help me. But I'm pretty desperate, and seeing I’m the only one who seems to have this problem, it's likely from something I installed during the perfect server setup. Does anything jump into any ones mind?
Thanks in advance.

Code:

 cat /proc/meminfo
MemTotal:        4069996 kB
MemFree:         1549640 kB
MemAvailable:     471476 kB
Buffers:          237056 kB
Cached:           637276 kB
SwapCached:       335536 kB
Active:          1262892 kB
Inactive:         892788 kB
Active(anon):     777464 kB
Inactive(anon):   572588 kB
Active(file):     485428 kB
Inactive(file):   320200 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:       2097148 kB
SwapFree:        1491196 kB
Dirty:                80 kB
Writeback:             0 kB
AnonPages:        992236 kB
Mapped:           234008 kB
Shmem:             68704 kB
Slab:             231556 kB
SReclaimable:     144120 kB
SUnreclaim:        87436 kB
KernelStack:        4352 kB
PageTables:        27296 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:     4132144 kB
Committed_AS:    3924732 kB
VmallocTotal:   34359738367 kB

 

clamd is the biggest in there at 10%, sql 5%, the reset is chickenfeed[slang = tiny]. I wish there was a similar command that showed who was putting how much in cache[can other software add to the cache, or can only the kernel use cache for diskio]? Something in there grows like a crazy about 10seconds after you start the machine. 5 seconds later the ram is full again. I've tried other things like
"echo 256000 > /proc/sys/vm/min_free_kbytes" and "memory pressure" that cache just keeps growing though.
total used free shared buff/cache available
Mem: 3974 1514 1414 90 1045 374
Swap: 2047 590 1457

Thanks again.

 

If it that fast, namely 15 seconds uptime and RAM is full, it probably is someting that starts at boot.
I have no idea what it could be, but I would try to find out by installing htop and dstat, and monitoring what is happening. That dstat has lots of info it can show, read the man page and choose those that are related to memory, swap etc. usage.
Use free -h to get readable output and post in CODE tags so columns stay readable.

 

Thanks for all your help, I think this is a bigger issue with azure dynamic memory. I've changed to static memory for now. doh.