NCSD Using Resources
On 1/28/17 I moved all of the sites on my server to run through HTTPS (thanks, Google). Everything seemed fine, but today I looked at my Munin report and see that I'm using a LOT of RAM! Like, 1/27/17 my peak was 3G (out of 4G), on the 28th my peak was 4.8G, and today I've been up to 7.5G. I'm currently at 6.28G.
I ran "top", and it shows that MySQL is the top user at 1366m (Virt), but NCSD is a close second at 1315m. The third thing in the list is rsyslogd at 243m.
I'm not surprised by MySQL... all of my sites use MySQL, and I'm not having a high number of access or anything, so that's probably normal. But NCSD is a new one for me. Since it's related to the domain, I'm only guessing... am I using more RAM now that I'm sending everything through HTTPS?
It's also notable that CPU usage has stayed virtually the same (around 0.8 peak), but Apache "number of threads" have gone up a lot; from a "regular" peak of around 275 to a peak today of 433 (currently, 404.27). Apache "accesses" is pretty consistent, but "processes" went from a peak of around 100 to a peak today of 184.
The only other thing I've changed is that I turned Keep-Alive back on, and set it to 3 seconds (it's been off for a few years). But I just did that last night, so while today is higher than yesterday, yesterday was still a lot higher than before I moved to HTTPS.
Is there a setting I should be changing in /etc/ncsd.conf? AFAIK, it's at the default right now. If not, would moving to PowerDNS help?
-
Hello, I've not seen any other reports where NSCD has resulted in performance issues. Could you open a support ticket using the link in my signature so we can take a closer look and see what's happening on your system? Thank you. 0 -
I started to submit a ticket, Michael, but the ticket center said for me to send it to my provider (Softlayer) first. I'll update the thread when/if there's a resolution. 0
Please sign in to leave a comment.
Comments
2 comments