AIX, Pains and Wines

The adidas Silverstone Half Marathon isn’t until 6 March 2011 but, I have received my Breakthrough Breast Cancer running vest already!

So, now it’s official. And that means it’s time to start raising money for Breakthrough Breast Cancer. Please visit my JustGiving page and donate as much as you can. Every donation counts, helping saving lives and changing futures and removing the fear of breast cancer for good.

Breakthrough Breast Cancer is a charity registered in England & Wales (No. 1062636) & Scotland (No. SC039058)

Having installed TSM 6.2.1 I ran the “Create Server Instance” wizard in the Administration Center. When attempting to run a database backup I got the following in the activity log:

ANR2017I Administrator ADMIN issued command: BACKUP DB dev=lto4 s=y (SESSION: 4)
ANR4559I Backup DB is in progress. (SESSION: 4)
ANR0984I Process 2 for DATABASE BACKUP started in the BACKGROUND at 15:37:37. (SESSION: 4, PROCESS: 2)
ANR2281I Incremental database backup started as process 2. (SESSION: 4, PROCESS: 2)
ANR0406I Session 6 started for node CCBKUAPPSP001 (DB2/AIX64) (Tcp/Ip loopback(57066)). (SESSION: 6)
ANR0422W Session 6 for node CCBKUAPPSP001 (DB2/AIX64) refused - node name not registered. (SESSION: 6)

It appears that TSM uses hard-coded values for the servername (TSMDBMGR_TSMINST1) and nodename ($$_TSMDBMGR_$$). The following lines were added to /usr/tivoli/tsm/client/api/bin64/dsm.sys but, not as the 1st (default) stanza:

servername TSMDBMGR_TSMINST1
commmethod tcpip
tcpserveraddr localhost
tcpport 1500
passwordaccess generate
passworddir /home/tsminst1/tsminst1
errorlogname /home/tsminst1/tsminst1/tsmdbmgr.log
nodename $$_TSMDBMGR_$$

The API password was reset by logging in as root and running the following:

# . /home/tsminst1/sqllib/db2profile
# /home/tsminst1/sqllib/adsm/dsmapipw

Specify TSMDBMGR as both the old and new password and restart the instance. Running the database backup now works as expected:

ANR2017I Administrator ADMIN issued command: BACKUP DB dev=lto4 s=y
ANR4559I Backup DB is in progress.
ANR0984I Process 6 for DATABASE BACKUP started in the BACKGROUND at 08:11:52.
ANR2281I Incremental database backup started as process 6.
ANR0406I Session 11 started for node $$_TSMDBMGR_$$ (DB2/AIX64) (Tcp/Ip loopback(64775)).
ANR8337I LTO volume C01000 mounted in drive RMT4 (/dev/rmt4).
ANR0511I Session 11 opened output volume C01000.
ANR1360I Output volume C01000 opened (sequence number 1).
ANR1361I Output volume C01000 closed.
ANR0514I Session 11 closed volume C01000.
ANR0403I Session 11 ended for node $$_TSMDBMGR_$$ (DB2/AIX64).
ANR0406I Session 12 started for node $$_TSMDBMGR_$$ (DB2/AIX64) (Tcp/Ip loopback(64784)).
ANR0511I Session 12 opened output volume C01000.
ANR1360I Output volume C01000 opened (sequence number 1).
ANR1361I Output volume C01000 closed.
ANR0514I Session 12 closed volume C01000.
ANR0403I Session 12 ended for node $$_TSMDBMGR_$$ (DB2/AIX64).
ANR4551I Incremental database backup (process 6) completed.
ANR0985I Process 6 for DATABASE BACKUP running in the BACKGROUND completed with completion state SUCCESS at 08:12:53.

During a recent network test, the both network switches in the primary computer-room were powered off. However, the PowerHA nodes (running as VIO clients) failed to detect a local adapter down event and the resource groups continued running on the active node.

This problem is explained more fully in the PowerHA for AIX Cookbook but, it is because the traffic being passed between the VIO clients looks like normal external traffic from the point of view of the LPAR’s OS. We were using the following netmon.cf (/usr/es/sbin/cluster/netmon.cf):

10.252.33.1
10.253.49.1
10.254.33.1
10.255.33.1

The solution is to use the new, intermediate, enhanced netmon.cf syntax:

!REQD owner target

!REQD is an explicit string that must be at the begining of the line, owner is the interface that uses this line and target is the IP address or hostname that the owner should try to ping. In a VIO configuration this should be an address outside of the VIO environment – our’s are the default gateways for the relevant networks. The example netmon.cf thus transforms into:

!REQD en0 10.252.33.1
!REQD en3 10.252.49.1
!REQD en2 10.254.33.1
!REQD en1 10.255.33.1

You may have noticed my use of the word, “intermediate” when describing this solution; the Cookbook was written for PowerHA v5.5 but, it appears to still apply for v6.2.1.

I spent a great Wednesday evening with my friends Dean and Marc who is Dean’s brother-in-law, American and also half-Hungarian. Marc Lives in Arlington, VA and has been over for a family wedding in Italy and the British Formula 1 Grand Prix. Dean lives about 1 mile from me (as the crow flies) in the West Midlands but, because of our work commitments, we see more of each other in London than anywhere else!

After a couple of pints we moved on to The Gay Hussar, an Hungarian restaurant in Greek Street, Soho. Marc had been before and a return trip can only be a good sign. The Maître’d (¼ Hungarian, ¾ Polish) suggested the Hungarian Hors d’Oeuvres would be a good choice to share. It turned out to be a very good recommendation, with a good mix of sausage, fish and salad. Whilst selecting our main courses, Marc chose a Pinot Noir from the wine list based purely on the fact that he recognised the name from a visit to Budapest 3 years ago!

Very soon we were tucking into healthy portions of Duck Livers Sautéed with Onions, Bacon and Paprika, Crispy Roast Duck with Red Cabbage, Hungarian Potatoes and Apple Sauce and Gypsy Quick Dish – Pork Medallions with Bacon, Onions, Potatoes and Paprika. The food went down very well. As did a second bottle of the wine. And, after a short break, so did dessert, although Dean needed a little help finishing his.

I’m not quite sure what had I expected of Hungarian cuisine but, I was was pleasantly surprised by the variety of choice, the generous portions and the excellent preparation of the food. You’d find it very hard to not find something you like.

I have entered my first race event, the adidas Silverstone Half Marathon on 6 March 2011. I’ll have to wait until October to see if my application for the Virgin London Marathon on 17 April 2011 has been successful.

IBM has identified a bug with version 6.2.1 of the TSM Backup and Archive Client. Using traditional scheduler services to manage a client backup, the backup reports the following error on a mounted filesystem:

ANS1071E Invalid domain name entered:/fs1

This is fixed in 6.2.2 but, the work-around for now is to use the Client Acceptor Daemon (/usr/bin/dsmcad) with the following added to your dsm.sys in the appropriate server stanza:

managedservices webclient schedule