Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. And see here for current known issues.

New Member

ACE Module uneven current connections

Good morning all,

  I currently have an issue where using the round-robin predictor we are getting some servers being hit with larger numbers of connections than others. This is causing the web servers to drop connections due to being hit very fast in a short period of time.

This is the output from a show serverfarm showing the issue. MK-HOST15 somehow has 1858 current connections and 48 connection failures ....

MK-ACE01/001# show serverfarm MK-FARM-sf
serverfarm     : MK-FARM-sf, type: HOST
total rservers : 8
---------------------------------
                                                ----------connections-----------
       real                  weight state        current    total      failures
   ---+---------------------+------+------------+----------+----------+---------
   rserver: MK-HOST10
       10.10.1.10:0          8      OPERATIONAL  92         206929     16
   rserver: MK-HOST11
       10.10.1.11:0          8      OPERATIONAL  93         206859     86
   rserver: MK-HOST12
       10.10.1.12:0          8      OPERATIONAL  93         206943     2
   rserver: MK-HOST13
       10.10.1.13:0          8      OPERATIONAL  116        206934     10
   rserver: MK-HOST14
       10.10.1.14:0          8      OPERATIONAL  93         206941     3
   rserver: MK-HOST15
       10.10.1.15:0          8      OPERATIONAL  1858       206896     48
   rserver: MK-HOST16
       10.10.1.16:0          8      OPERATIONAL  93         206935     9
   rserver: MK-HOST17
       10.10.1.17:0          8      OPERATIONAL  95         206838     106

When this happens the failure counter for MK-HOST15 increases a lot

MK-ACE01/001# show serverfarm MK-FARM-sf
serverfarm     : MK-FARM-sf, type: HOST
total rservers : 8
---------------------------------
                                                ----------connections-----------
       real                  weight state        current    total      failures
   ---+---------------------+------+------------+----------+----------+---------
   rserver: MK-HOST10
       10.10.1.10:0          8      OPERATIONAL  229        345394     80
   rserver: MK-HOST11
       10.10.1.11:0          8      OPERATIONAL  216        345347     126
   rserver: MK-HOST12
       10.10.1.12:0          8      OPERATIONAL  210        345275     201
   rserver: MK-HOST13
       10.10.1.13:0          8      OPERATIONAL  669        345369     106
   rserver: MK-HOST14
       10.10.1.14:0          8      OPERATIONAL  210        345442     32
   rserver: MK-HOST15
       10.10.1.15:0          8      OPERATIONAL  203        345167     309
   rserver: MK-HOST16
       10.10.1.16:0          8      OPERATIONAL  228        345417     59
   rserver: MK-HOST17
       10.10.1.17:0          8      OPERATIONAL  256        345370     108

This behavior keeps occuring and I am not able to figure out why the current connections shows uneven numbers to such a large degree. The overall connection counters seem approximately even though.

We are running ACE Code A2(2.3)

The web servers are running Apache and keepalives are off inline with our production enviroments. TCP-Reuse is also turned off. There is no stickyness configured.

Any ideas why certain servers keep getting hit with more connections than others ? It's like roundrobin keeps stopping at a server for a second before moving on to the next one. Any suggestions appreciated.

Thanks.

318
Views
0
Helpful
0
Replies
CreatePlease to create content