To answer the first question, asynchronous replication (to the second storage device) is operated by a queue sweep, which can be tuned/scheduled in various ways in common with all queue sweeps (refer to the documentation). However, in the case of a down connection a couple of additional factors apply:
- in general within the advanced storage logic there is a delay between attempts to re-restablish a connection to a storage device which was unavailable which defaults to 5 minutes. This can be overridden through the JVM argument Advanced.ConnectivityRefreshSeconds.
- advanced storage queue items which cannot be processed due to a down connection are subject to deferral for a period which defaults to two minutes beyond the connectivity refresh period. This can also be overridden by the JVM argument Advanced.Queue.Entry.Defer.Seconds (which would be set to an absolute time in seconds).
The idea behind this is to avoid thrashing and unnecessary failure-retry of queue items, and we don't recommend changing either setting.
Regarding the second question, Gerold already somewhat answered. But to be very clear, bulk move can move from
any type of storage area to
any other type.
------------------------------
Mike
------------------------------
Original Message:
Sent: Sun January 10, 2021 03:24 AM
From: Christos Chorattides
Subject: Looking for HA fileshare solution for Content Engine on Windows ND cluster
What a great idea Michael! Thanks!.
I have never used Advanced Storage Areas before. I knew they had something to do with Cloud Storage and Replication but I didn't think I could use with a WAS Cluster to have a highly available storage.
I tested two areas-two devises and seems to wok fine! Have you used this in a production environment?
I have a few questions, if you know. When I tested with the second Storage Area down, system was operating fine. After bringing up the second storage, it took some time to synchronize. What is the default interval in which the first storage area checks and synchronizes the second storage area after it comes up? Is it configurable?
Also on "normal" Storage Areas we can add multiple Storage Areas to the same Storage Policy, creating a storage farm to improve throughput and implement size and file number constrains for each Storage Area. Can we implement something similar to Advanced Storage Areas?
Can the Bulk Content Move Sweeps move content from "normal" Storage Areas to Advanced Storage Areas?
BR
------------------------------
Christos Chorattides
Datatech
Original Message:
Sent: Fri January 08, 2021 11:51 AM
From: Michael Seaman
Subject: Looking for HA fileshare solution for Content Engine on Windows ND cluster
Have you considered using an Advanced Storage Area with two file system storage devices?
See https://www.ibm.com/support/knowledgecenter/SSNW2F_5.5.0/com.ibm.p8.planprepare.doc/p8ppi267.htm
------------------------------
Mike
Original Message:
Sent: Thu December 31, 2020 02:45 AM
From: Christos Chorattides
Subject: Looking for HA fileshare solution for Content Engine on Windows ND cluster
Hello everyone.
we have been successfully deploying Content Engine with WAS ND clusters on AIX using GPFS as a shared highly available filesystem.
Now we need to deploy CE with WAS ND on Windows 2019 and we are looking for the most efficient, easy to configure (no third party licensed products pls), shared-higly available filesystem for CE filestores.
Our Windows Servers are VMs on VMware infrastructure.
Any ideas from colleagues that use Windows servers?
BR
Christos
------------------------------
Christos Chorattides
Datatech
------------------------------