New answers tagged

0

The answer according to me is that network bandwidth gets used when heavy memory is consumed by an application (which has 0 affect on UI). This might not sound right to many IT professionals but in my case when i tried to reduce the heavy memory consumption by removing images from the memory automatically the network bandwidth consumption reduced. ...


0

Not directly, no. If you can get a packet capture on the network utilization during the calculation, you'll see what's using the bandwidth. Perhaps the server has to grab a bunch of data from a backend DB box to perform said calculation - that'd be my guess.


0

The issue with django and memcached is that it makes a connection each request. So part of that time is connection setup time. It depends on which memcache binding you are using. But you could put something like this in your wsgi.py # Fix django closing connection to memcached after every request (#11331) from django.core.cache.backends.memcached import ...


1

You are not wasting space. tmpfs has to be used before it will use RAM and paging space. free -h will show you memory use in human readable units. If you were using tmpfs space, that would be under shared.


0

I've just worked through this error on a Lenovo NAS running Debian 5 and kernel 2.6.39.3 64bit. The messages are informational despite looking scary, according to https://www.novell.com/support/kb/doc.php?id=7002803 However they were filling my very limited root partition (this device has a 50 MByte root partition ?!) The fix for me was to set vm....


1

Depending on the number of your server's cores follow the table in this link For two processors (probably your case) you should place 6 DIMMs in A1 to A6 6 DIMMs in B1 to B6


0

I could now verify what the problem was in my case: I use a software, which is able to delete its own caches and which uses a package system, where you can install new packages. Whenever I install a new package or clear the cache, the software runs a opcache_reset() instead of just using opcache_invalidate() because it knows which files are about to be ...


0

Only DDR4 Reg ECC RAM and Load-Reduced Reg ECC RAM will work. Refer this link for compatible memory upgrades for Supermicro X10SRL-F Motherboard http://www.memorystock.com/memory/SuperMicroX10SRLF.html


4

No, a NUMA system is a NUMA system, there are nodes (generally CPUs), each node has it's own memory - most modern OS's will try to ensure that all memory allocations to a process are on the same node that the process is running. If this can't happen then yes you get the memory slowdown you've seen where a node is essentially acting like a memory controller ...


0

Here is my hacky version. Ideas how to improve it, are welcome: List the amount of RAM in kiB every VM needs: hypervisor5:~ # for dom in $(virsh list --all --name); do echo $(virsh dumpxml $dom | sed -nre 's/^.*>([0-9]*)<.currentMemory.*$/\1/p') $dom ; done| sort -n 524288 testfoo 2146304 x131 3121152 y114 3121152 foo 4096000 y123 4194304 mac-test ...



Top 50 recent answers are included