Update: SAP HANA Developer Edition v1.8.80
The following is an update regarding the HANA Developer Edition. For those who are currently running a system the following SAP Note is critical for you to read.
http://service.sap.com/sap/support/notes/1888072
There is a bug in the Linux glibc library that leads to an invalid memory access. SLES 11 SP 2 and SLES 11 SP 3 are affected. This could very well effect every query you attempt or try to run. So it’s certainly best that you make the updates prescribed in the note.
You will need to follow the steps in the SAP note to fix the problem with the underlying server itself.
We are working on a fix ourselves with the SAP Cloud Appliance Library so that the fix is applied automatically when you get a version of the system however for those who have already gotten a system you will need to manually fix yours.
Hi Craig,
is CAL fixed automatically already or is it going to be for the before "Activate" step for the server cloud instance ?
rgds,
greg
I am speaking to the team in the morning, the version should at the moment be no longer visible as of today the 18th. Once the fix is in place it will be visible again - anyone prior to that point would have to do the fix manually.
We just became aware of all of this so we are unfortunately reacting at the moment so chances are things are still in motion.
Thanks for the info Craig -
Unfortunately I don't have access to the patch through the SAP Note link to SUSE. I was able to update my instance using the command (run as root) "zypper update" to update all the packages on the SUSE server through ssh..
It's not a patch, just instructions I'm asking now if I can publish the content here as well since some of you may not have access.
The RPM packages libgcc_s1 and libstdc++6 need to be installed, please ensure to run at least the following versions:
libgcc_s1-4.7.2_20130108-0.17.2
libstdc++6-4.7.2_20130108-0.17.2
Hi Craig,
i apologize for being a bit dense here, but if i activate cal now, i should not be worrying about those RPM packages as i'm not sure if aws would let me do anything with them anyway.
is this a fair assumption?
thx,
greg
I would not activate it just yet. In fact you should not be able to actually see it at the moment.
are we talking about the same instance or has it just become available?
SAP HANA, developer edition 1.0 Rev 80
Description: Invent new possibilities with SAP HANA, a completely re-imagined, modern platform for real-time business: Run your business in real real time. SAP and partner solutions powered by SAP HANA can help you dramatically accelerate analytics, business processes, predictive analysis, and sentiment data processing – all on a single in-memory computing platform. Important: please observe that usage of this virtual appliance is governed by Terms and Conditions as published in the Pricing section. In particular 3c Publisher: SAP AG
Available In: Amazon Web Services, Microsoft Azure Update: 19 (Jun 23 2014, 03:03:16 GMT-04:00)
Virtual Machines Count: 1 Recommended VM Sizes: m2.4xlarge (Amazon Web Services)
I will check on that, I was not informed if they had a fix yet or not - however it's easy to update as I described earlier in another comment.
I tried the same - but now my instance is not reachably any more.
zypper update did nothing, so I executed sces-activate before. Now zypper update made a lot of updates. libgcc_s1 was updated, too.
"zypper ps" then showed me that there are some processes which need to be restarted because they are using the old versions. This included hdbindexserver, hdbdaemon etc....
I restarted HANA Database via Studio - but some services have still been running. So I decided to restart the machine via EC2 Console.
It was running at first (and zypper ps wasn't showing any processes any more), but then a few minutes later the system is not reachable any more. EC2 Console shows me Instance reachability check failed 🙁
When I view the system log the last lines are
[m [?25hMaster Resource Control: runlevel 3 has been
[80C [10D [1mreached [m
Failed services in runlevel 3: [80C [26D [1;31mnetwork sces-client [m
Skipped services in runlevel 3: [80C [24D [1;33mnfs mcelog smartd [m
Welcome to SUSE Linux Enterprise Server 11 SP3 (x86_64) - Kernel 3.0.101-0.8-ec2 (xvc0).
sid-hdb login:
but before there are a few "FATAL" messages:
[m [?25hLoading required kernel modules
7 [?25l [1A
[80C [10D [1;32mdone [m 8 [?25h [m [?25hStarting MD RAID FATAL: Could not load /lib/modules/3.0.101-0.8-ec2/modules.dep: No such file or directory
... no MD support in kernel 7 [?25l
[80C [10D [1;33mskipped [m 8 [?25h
...
System Boot Control: Running /etc/init.d/boot.local
FATAL: Could not load /lib/modules/3.0.101-0.8-ec2/modules.dep: No such file or directory
iptables-restore v1.4.6: iptables-restore: unable to initialize table 'nat'
Error occurred at line: 2
Try `iptables-restore -h' or 'iptables-restore --help' for more information.
7 [?25l [1A
[80C [10D [1;31mfailed [m 8 [?25h
Is there anything I can do to recover the machine? Or should I set up a new image and restore a data backup?
Any hints appreciated!
Another interesting part:
[m [?25hSetting up (localfs) network interfaces:
lo name: 0
lo IP address: 127.0.0.1/8
IP address: 127.0.0.2/8
7 [?25l [1A
[80C [10D [1;32mdone [m 8 [?25h eth0 name: 0
FATAL: Could not load /lib/modules/3.0.101-0.8-ec2/modules.dep: No such file or directory
eth0 Starting DHCP4 client. . . . . . . .
eth0 DHCP4 continues in background
[1A
[80C [10D [1;33mwaiting [m
Waiting for mandatory devices: eth0
12 11 9 8 7 6 5 3 2 1 0
eth0 name: 0
eth0 DHCP4 client (dhcpcd) is running
eth0 . . . but is still waiting for data
eth0 IP address: 172.31.24.42/20
eth0 is up
[1A
[80C [10D [1;33mwaiting [m
eth0 interface could not be set up until now
7 [?25l [1A
[80C [10D [1;31mfailed [m 8 [?25hSetting up service (localfs) network . . . . . . . . . . 7 [?25l
[80C [10D [1;31mfailed [m 8 [?25h
[m [?25hStart applianceagent serviceFATAL: Could not load /lib/modules/3.0.101-0.8-ec2/modules.dep: No such file or directory
iptables v1.4.6: can't initialize iptables table `nat': iptables who? (do you need to insmod?)
Perhaps iptables or your kernel needs to be upgraded.
Sorry - I should have updated my status on this. I also lost connectivity to the machine after I rebooted it. Since I had not imported my Delivery Unit, I decided I would terminate the instance and wait for a more stable release. Sorry for any issues this might have caused you.
I think you could just update glibc rpm/packages rather than the entire machine if you still have the older version.
Jim
Hey Jim,
thanks for your reply. I terminated my instance too, set up a new one and restored backup. So it seems to be confirmed that zypper update is a bad idea!
Just wanted to give this warning before anybody other gets in trouble, too...
Many thanks for keeping the HANA developer community updated Craig, appreciated.
Hi Craig, with the availability of Rev 80 showing on CAL, grateful for confirmation we good to go now with creation of instance, with fix in place. Cheers.