[Freeipa-users] Excessive CPU usage by ns-slapd

[Jatin Nansi]

Check the ns-slapd access and error logs of the DS instance hosting the 
IPA instance. The strace output indicates that ns-slapd spent most of 
its time waiting for network activity to happen (poll), which is normal 
for ns-slapd.

Jatin
On 19/02/15 15:52, Jim Richard wrote:
> I’ve got 4 Redhat IDM masters in a multi-master 
> config. 3.0.0-42.el6.centos is the IPA version, 389-ds-base version 
> 1.2.11.15-48.el6_6, Centos 6.6
>
> Monitoring established connections on port 389 and dsInOps over time 
> shows a consistent/even level of activity however 2 of the 4 IPA 
> servers show ever increasing CPI usage by ns-slapd. One ns-slapd 
> process will start to show increased CPU for a time, then drop off as 
> another then increases. This cycle continues with each switch seeing 
> more and more total SPU usage by ns-slapd.
>
> strace timing for the offending ns-slapd looks like the following:
>
>
> % time     seconds  usecs/call     calls    errors syscall
> ------ ----------- ----------- --------- --------- ----------------
>  96.12    9.342272        1133      8243 poll
>   3.86    0.375457          53      7066        41 futex
>   0.01    0.000668           0      8244      8244 getpeername
>   0.00    0.000374           0       929 close
>   0.00    0.000368           0      3201 read
>   0.00    0.000151           0       882 setsockopt
>   0.00    0.000095           2        42 access
>   0.00    0.000033           0      1365 fcntl
>   0.00    0.000000           0        42 open
>   0.00    0.000000           0        39 stat
>   0.00    0.000000           0        42 fstat
>   0.00    0.000000           0         1 madvise
>   0.00    0.000000           0       441 accept
>   0.00    0.000000           0       441 getsockname
>   0.00    0.000000           0         1 restart_syscall
> ------ ----------- ----------- --------- --------- ----------------
> 100.00    9.719418                 30979      8285 total
>
>
> I have carefully reviewed cn=config settings on all four master 
> servers to confirm that they match.
>
> Based on this strace output can you perhaps point me in the right 
> direction, give me a clue on what I should be looking at.
>
> Here’s a screen shot of my Zabbix reporting to help describe the 
> problem. Note the graph in the bottom right corner.
>
> The problem is most certainly related to replication but I just don’t 
> know what specifically to look at.
>
>
>
>
> Thanks in advance for any clues you can provide.
>
>
>
>
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> 	
> Jim Richard  | PlaceIQ 
> <http://www.google.com/url?q=http%3A%2F%2Fwww.placeiq.com%2F&sa=D&sntz=1&usg=AFrqEzcYjZpDPyqW7feNK9EgLq-c9JlHiw>  | 
>  Systems Administrator  |  jrichard at placeiq.com 
> <mailto:name at placeiq.com>  | +1 (646) 338-8905
>
>
>
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/freeipa-users/attachments/20150219/f74fc50f/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/png
Size: 1563895 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/freeipa-users/attachments/20150219/f74fc50f/attachment.png>