[arch-general] mysqld (MySQL and MariaDB) polling
Hello everybody, based on the thread "[RFC] Migration to MariaDB" in arch-dev-public by Bartłomiej Piotrowski I started playing with MySQL and MariaDB. I noticed both were polling every second: <... futex resumed> ) = -1 ETIMEDOUT (Connection timed out) futex(0x1d48050, FUTEX_WAKE_PRIVATE, 1) = 0 gettimeofday({1361806174, 246573}, NULL) = 0 futex(0x1d4808c, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 2849, {1361806175, 246573000}, ffffffff <unfinished ...> Searching Google I found that the problem may be caused by the leap second, though the load is not as high as some articles describe. (Continues polling vs. one poll a second.) However the polling stops as soon as I manually set the time with date -s and it start again if I run ntpdate to sync the time. As this is a notebook I rebooted more than once since the leap second happened. ;) What is going on? This is a mobile system, so it would be great to get rid of mysqld eating my battery power. -- main(a){char*c=/* Schoene Gruesse */"B?IJj;MEH" "CX:;",b;for(a/* Chris get my mail address: */=0;b=c[a++];) putchar(b-1/(/* gcc -o sig sig.c && ./sig */b/42*2-3)*42);}
On Mon, Feb 25, 2013 at 4:40 PM, Christian Hesse <list@eworm.de> wrote:
based on the thread "[RFC] Migration to MariaDB" in arch-dev-public by Bartłomiej Piotrowski I started playing with MySQL and MariaDB. I noticed both were polling every second:
<... futex resumed> ) = -1 ETIMEDOUT (Connection timed out) futex(0x1d48050, FUTEX_WAKE_PRIVATE, 1) = 0 gettimeofday({1361806174, 246573}, NULL) = 0 futex(0x1d4808c, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 2849, {1361806175, 246573000}, ffffffff <unfinished ...>
Searching Google I found that the problem may be caused by the leap second, though the load is not as high as some articles describe. (Continues polling vs. one poll a second.)
However the polling stops as soon as I manually set the time with date -s and it start again if I run ntpdate to sync the time. As this is a notebook I rebooted more than once since the leap second happened. ;)
What is going on? This is a mobile system, so it would be great to get rid of mysqld eating my battery power.
Related: https://plus.google.com/114657443111661859546/posts/VSrk6BfLnNe ? I always thought this was "just the way it is", but maybe you are onto something... Might be worth reporting upstream. -t
On 25 Feb 2013 15:40, "Christian Hesse" <list@eworm.de> wrote:
Hello everybody,
based on the thread "[RFC] Migration to MariaDB" in arch-dev-public by Bartłomiej Piotrowski I started playing with MySQL and MariaDB. I noticed both were polling every second:
<... futex resumed> ) = -1 ETIMEDOUT (Connection timed out) futex(0x1d48050, FUTEX_WAKE_PRIVATE, 1) = 0 gettimeofday({1361806174, 246573}, NULL) = 0 futex(0x1d4808c, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 2849, {1361806175, 246573000}, ffffffff <unfinished ...>
Searching Google I found that the problem may be caused by the leap
though the load is not as high as some articles describe. (Continues
second, polling
vs. one poll a second.) Hello
I had the same problem with an EC2 instance which cost me some pounds.. And yeah from what I searched as well it probably relevant to the leap second problem.
However the polling stops as soon as I manually set the time with date -s
and
it start again if I run ntpdate to sync the time. As this is a notebook I rebooted more than once since the leap second happened. ;)
Mine was fixed when I did the date -s. (And I have done reboot after it.) Have you tried changing servers for the ntpd? I think there was a bug report in fedora about it, maybe worth looking for it and check comments. Also do you update the hwclock?
What is going on? This is a mobile system, so it would be great to get
rid of
mysqld eating my battery power. -- main(a){char*c=/* Schoene Gruesse */"B?IJj;MEH" "CX:;",b;for(a/* Chris get my mail address: */=0;b=c[a++];) putchar(b-1/(/* gcc -o sig sig.c && ./sig */b/42*2-3)*42);}
participants (3)
-
Christian Hesse
-
Leonidas Spyropoulos
-
Tom Gundersen