Last week at IMPACT in Las Vegas we announced IBM MessageSight, an appliance that can be placed in a DMZ that handles outbound & inbound messages at scale. While primarily useful for machine-to-machine (m2m) communication using MQTT it also handles WebSocket connections for web applications.
While there are quite a number of communication options an important concept to understand is that this appliance is not a queue manager or broker. It’s basically a messaging gateway that makes it possible to only connect a small number of devices to internal IT systems, but still gather or push messages to millions of devices. A full list of its key capabilities can be found in the announcement letter.
For analyzing the CMS collector of Oracle JVMs the posts by Jon Masamitsu about the Phases of CMS and its PrintGCDetails Output are very useful. I’d also recommend to look at the Java HotSpot Garbage Collection documentation, but I’ve not found it very comprehensive in the past and not as easy to use as our IBM Java Diagnostics Guide.
Also be aware that some of the behavior you’re seeing in verbose gc output might be because of the characteristics of the underlying platform. I’ve seen JVMs with large amounts of threads take a long time to be stopped by the collector on Solaris on Niagara boxes, so don’t assume different behavior between platforms is necessarily a defect in the GC itself.
An updated WebSphere Application Server V8.5.Next Beta is now available. The Liberty profile now includes a JMS messaging engine with support for MDBs, enhanced CDI support, federated LDAP registry support, the ability to act as an OAuth Service Provider, the ability to organize Liberty servers into collectives for clustering and administration, and more.
Internally at IBM we use BlueGroups to manage permissions for various services. It’s basically just regular LDAP groups inside of a Tivoli Directory Server. We have a web UI for managing those groups. Unfortunately the definition of managing apparently doesn’t include finding the list of groups you’re a member of. For this you need some ldapsearch trickery.
First, find the
dn for your user:
$ ldapsearch -x -h bluepages.ibm.com -b ou=bluepages,o=ibm.com "(mail=NAME@CC.ibm.com)" dn
Then use that
dn to find the groups you’re a member of:
$ ldapsearch -x -h bluepages.ibm.com -b ou=memberList,ou=ibmgroups,o=ibm.com "(uniqueMember=uid=SSSSSSCCC,c=CC,ou=bluepages,o=ibm.com)" dn
Working as a middleware admin? Then you should really check out t-rob’s latest post. This post works for any middleware you’re working with, not only MQ :-)
Want some more insight into the magic that goes on inside of WebSphere Application Server Network Deployment’s dynamic clusters? Then check out the paper about A Scalable Application Placement Controller for Enterprise Data Centers.
IBM United States Software Announcement 212-417:
IBM extends the end-of-service date for IBM WebSphere Application Server V6.1 from September 30, 2012, to September 30, 2013.
Jigsaw will most likely be delayed again. This time until Java 9. The goals are ambitious, but given a first demo of this was shown at JavaOne three years ago, this is pretty disappointing. You can find the FAQ on Mark’s blog and join the discussion there or send your feedback directly to the Java SE 8 Comments List.
Amazon yesterday released their latest Amazon Web Service called Amazon Glacier, a remote archiving/long-term storage service. The downside is that Amazon can take several hours to make your data available for download after you request it, but that’s not too bad when used as an archive or backup. You also can’t arbitrarily name objects stored in Amazon Glacier, as objects are named by the service when you store data there, so you need to remember which object is which after you store it.
Apart from the obvious use as a remote backup, which is kind of boring, this might be interesting for use as an archive storage to store things “just in case” for a limited period of time. I got really used to the Tivoli Storage Manager archive feature on servers. You just put something in the archive that you would rather delete, but don’t know if you might need it back in the next few months. It gets stored in the archive for later retrieval and if you don’t need it anymore it just vanishes from the archive after 90 days, or whatever timeout you set when storing it in the archive.
Wouldn’t it be awesome to have something like this without the hassle of having your own tape library? Amazon’s cost model even seems to favor this use. You don’t pay for incoming data transfer, monthly storage costs are rather low, and you only pay a penalty if you remove data earlier than 90 days.