[jira] [Comment Edited] (INFRA-11193) ldaps://ldap2-us-west.apache.org:636 often does not respond or times out when searching

["Sebb (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/INFRA-11193?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15197794#comment-15197794 ] 

Sebb edited comment on INFRA-11193 at 3/16/16 6:15 PM:
-------------------------------------------------------

FTR

This is the main query that fails when run by Whimsy:

ldapsearch -x -LLL -H ldaps://ldap2-us-west.apache.org:636/ -b ou=people,dc=apache,dc=org -s one '(|(cn=*)(loginShell=*)(asf-personalURL=*)(createTimestamp=*)(modifyTimestamp=*))' cn loginShell asf-personalURL createTimestamp modifyTimestamp uid

This particular search failed with Time Limit exceeded at the following times so far in March:
[Tue Mar 15 11:06:42 2016]
[Tue Mar 15 09:06:43 2016]
[Tue Mar 15 09:50:43 2016]
[Mon Mar 14 11:20:39 2016]
[Mon Mar 14 12:06:39 2016]
[Wed Mar  9 09:50:42 2016]
[Wed Mar  9 14:35:56 2016]
[Wed Mar  9 16:50:42 2016]
[Fri Mar  4 09:06:03 2016]
[Fri Mar  4 10:36:25 2016]
[Thu Mar  3 11:06:00 2016]
[Thu Mar  3 12:05:42 2016]
[Wed Mar  2 11:35:53 2016]

It does not occur every time, but has happened more than once, so is clearly reproducible.

The queries run 4 times an hour, so the failure rate is 2-3% on a day when it fails.
Overall, the failure rate is more like 1%.

We have rarely seen Time Limit exceeded for other queries such as:
ldapsearch -x -LLL -b ou=pmc,ou=committees,ou=groups,dc=apache,dc=org -s one cn=* dn member modifyTimestamp
and
ldapsearch -x -LLL -b ou=groups,dc=apache,dc=org -s one cn=* dn memberUid modifyTimestamp
but these return less data.

Note that we have only seen it for ldap2-us-west and occasionally for ldap-lb-us.
[That appears to be a load balancing URL so may well have actually targeted ldap2-us-west]

As far as the Whimsy code is concerned, so far it is definitely reproducible.

I'm not sure about verification, short of running the query repeatedly every 15 mins or so for several days. I could set up such a cron job, but it seems wasteful just to repeat what Whimsy is already doing.


was (Author: sebb@apache.org):
This is the query that fails when run by Whimsy:

ldapsearch -x -LLL -H ldaps://ldap2-us-west.apache.org:636/ -b ou=people,dc=apache,dc=org -s one '(|(cn=*)(loginShell=*)(asf-personalURL=*)(createTimestamp=*)(modifyTimestamp=*))' cn loginShell asf-personalURL createTimestamp modifyTimestamp uid >x


> ldaps://ldap2-us-west.apache.org:636 often does not respond or times out when searching
> ---------------------------------------------------------------------------------------
>
>                 Key: INFRA-11193
>                 URL: https://issues.apache.org/jira/browse/INFRA-11193
>             Project: Infrastructure
>          Issue Type: Bug
>          Components: LDAP
>            Reporter: Sebb
>            Assignee: Chris Lambertus
>            Priority: Trivial
>         Attachments: PMBLdap.txt
>
>
> The LDAP server ldaps://ldap2-us-west.apache.org:636 often fails to respond when contacted by Whimsy.
> Note that the server does usually work OK. However AFAICR it is the only one that has been reported as failing by Whimsy, so there may be an issue with the server or connectivity that is worth noting.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)