Mailing-List: contact users-help@directory.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@directory.apache.org
From: "Accorsi, Carlo" <Carlo.Accorsi@siemens.com>
To: "users@directory.apache.org" <users@directory.apache.org>
Subject: Question about I/O and VM failover
Thread-Topic: Question about I/O and VM failover
Thread-Index: AdFoEjXnKOBDUH19RziZ4CsuVoy7qA==
Date: Mon, 15 Feb 2016 17:14:07 +0000
Message-ID: 
 <FC9D57306008D54AA73E719FC154816647F11190@USSLMMBX003.net.plm.eds.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: multipart/alternative;
	boundary="_000_FC9D57306008D54AA73E719FC154816647F11190USSLMMBX003netp_"
MIME-Version: 1.0

--_000_FC9D57306008D54AA73E719FC154816647F11190USSLMMBX003netp_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Hi,

We had a customer a couple of weeks ago call us and the JDBM store kept get=
ting corrupted. We'd rebuild the directory database from scratch and a coup=
le days later it would
become corrupted again and would not start. It then became known that the V=
irtual Machine ApacheDS was running on was failing over between two several=
 times a day. In simple terms, when a virtual machine fails over, the hyper=
visor suspends all I/O to the box, then moves it to another host (hardware)=
 then restores the I/O when it comes online again with the same IP.

So my question is this. This failover process can sometimes take cases 10 o=
r 20 seconds to complete. During this time any thread trying to write to th=
e filesystem cannot. This is likely some sort of hardware problem but it go=
t us wondering if there are any special considerations for this type of sce=
nario?  For example, are there configurations for thread write timeout?
For now, the machine is not failing over and everything is ok.

Many Thanks!!
Carlo

--_000_FC9D57306008D54AA73E719FC154816647F11190USSLMMBX003netp_--