Tell me more ×
Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It's 100% free, no registration required.
up vote 2 down vote favorite

For the last few days I've been trying to figure out why my MySQL read slave servers are capped at 1012+1 (+1 for the superuser) connections when I have the max_connections set to 5000. It seems that no matter what I change, the "Historic max_used_connections" never increases beyond 1013.

The mysql read slaves support a high traffic web site. I have a cluster of 3 read only slaves. Each read slave seems to be capped at 1013 max connections. I can take two out of the cluster so all connections go to one and still only 1013. No error messages show up in the mysql log file.

A mysqladmin status on one of the read slaves shows this:
Uptime: 1661 Threads: 128 Questions: 17634224 Slow queries: 0 Opens: 2441 Flush tables: 1 Open tables: 938 Queries per second avg: 10616.630

Same server, mtop shows this:
load average: 0.05, 0.15, 0.19 mysqld 5.1.66-log up 0 day(s), 0:28 hrs 106 threads: 4 running, 906 cached. Queries/slow: 1.6K/0 Cache Hit: 99.96% Opened tables: 0 RRN: 278 TLW: 1.0K SFJ: 0 SMP: 0 QPS: 1

Each server has 12 dual-core Intel CPU's (24 cores) with 140 GB of RAM running RHEL 6.3, Kernel 2.6. (x86_64)

Any idea what would be limiting/capping the connections at 1012+1 ?

EDIT Also wanted to add two things:

1) I am running the "free/community" edition of MySQL that you get from a Red Hat 6.3 YUM install. Could there be some limitation in that version?

2) I've changed the /etc/security/limits.conf file on the MySQL read slaves, increasing values for users mysql and root (did not help my situation):

core file size          (blocks, -c) 0
data seg size           (kbytes, -d) unlimited
scheduling priority             (-e) 0
file size               (blocks, -f) unlimited
pending signals                 (-i) 1161534
max locked memory       (kbytes, -l) 64
max memory size         (kbytes, -m) unlimited
open files                      (-n) 10240
pipe size            (512 bytes, -p) 8
POSIX message queues     (bytes, -q) 819200
real-time priority              (-r) 0
stack size              (kbytes, -s) 10240
cpu time               (seconds, -t) unlimited
max user processes              (-u) 32768
virtual memory          (kbytes, -v) unlimited
file locks                      (-x) unlimited
share|improve this question
client-side (e.g. php, apache maxchilds) limits lower than the server limits? – Marc B Jan 15 at 21:14
2  
Not sure if this is relevant, but could it be the file descriptor limit? – Patrickdev Jan 15 at 21:15
1  
1013 sounds suspiciously close to the default 1024 file handle limit imposed on most operating systems. – tadman Jan 15 at 21:24
@MarcB no, according to our Apache configuration we could potentially have a total (all Apache servers that connect to the db's) of 8400 connections. – Dan Arnold Jan 15 at 21:39
1  
Remember ulimit only applies to that shell and doesn't affect other processes. – tadman Jan 15 at 21:42
show 7 more comments

migrated from stackoverflow.com Jan 16 at 1:08

Know someone who can answer? Share a link to this question via email, Google+, Twitter, or Facebook.

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.