Best practices for ZFS file systems when using live upgrade?

I would like feedback on how to layout the ZFS file system to deal with files that are constantly changing during the Live Upgrade process. For the rest of this post, lets assume I am building a very active FreeRadius server with log files that are constantly updating and must be preserved in any boot environment during the LU process.
Here is the ZFS layout I have come up with (swap, home, etc omitted):
NAME                                USED AVAIL REFER MOUNTPOINT
rpool                              11.0G 52.0G    94K /rpool
rpool/ROOT                         4.80G 52.0G    18K legacy
rpool/ROOT/boot1                   4.80G 52.0G 4.28G /
rpool/ROOT/boot1/zones-root         534M 52.0G    20K /zones-root
rpool/ROOT/boot1/zones-root/zone1   534M 52.0G   534M /zones-root/zone1
rpool/zone-data                      37K 52.0G    19K /zones-data
rpool/zone-data/zone1-runtime        18K 52.0G    18K /zones-data/zone1-runtimeThere are 2 key components here:
1) The ROOT file system - This stores the / file systems of the local and global zones.
2) The zone-data file system - This stores the data that will be changing within the local zones.
Here is the configuration for the zone itself:
<zone name="zone1" zonepath="/zones-root/zone1" autoboot="true" bootargs="-m verbose">
<inherited-pkg-dir directory="/lib"/>
<inherited-pkg-dir directory="/platform"/>
<inherited-pkg-dir directory="/sbin"/>
<inherited-pkg-dir directory="/usr"/>
<filesystem special="/zones-data/zone1-runtime" directory="/runtime" type="lofs"/>
<network address="192.168.0.1" physical="e1000g0"/>
</zone>The key components here are:
1) The local zone / is shared in the same file system as global zone /
2) The /runtime file system in the local zone is stored outside of the global rpool/ROOT file system in order to maintain data that changes across the live upgrade boot environments.
The system (local and global zone) will operate like this:
The global zone is used to manage zones only.
Application software that has constantly changing data will be installed in the /runtime directory within the local zone. For example, FreeRadius will be installed in: /runtime/freeradius
During a live upgrade the / file system in both the local and global zones will get updated, while /runtime is mounted untouched in whatever boot environment that is loaded.
Does this make sense? Is there a better way to accomplish what I am looking for? This this setup going to cause any problems?
What I would really like is to not have to worry about any of this and just install the application software where ever the software supplier sets it defaults to. It would be great if this system somehow magically knows to leave my changing data alone across boot environments.
Thanks in advance for your feedback!
--Jason

Hello "jemurray".
Have you read this document? (page 198)
http://docs.sun.com/app/docs/doc/820-7013?l=en
Then the solution is:
01.- Create an alternate boot enviroment
a.- In a new rpool
b.- In the same rpool
02.- Upgrade this new enviroment
03.- Then I've seen that you have the "radious-zone" in a sparse zone (it's that right??) so, when you update the alternate boot enviroment you will (at the same time) upgrading the "radious-zone".
This maybe sound easy but you should be carefull, please try this in a development enviroment
Good luck

Similar Messages

How to delete file systems from a Live Upgrade environment

How to delete non-critical file systems from a Live Upgrade boot environment?
Here is the situation.
I have a Sol 10 upd 3 machine with 3 disks which I intend to upgrade to Sol 10 upd 6.
Current layout
Disk 0: 16 GB:
/dev/dsk/c0t0d0s0 1.9G /
/dev/dsk/c0t0d0s1 692M /usr/openwin
/dev/dsk/c0t0d0s3 7.7G /var
/dev/dsk/c0t0d0s4 3.9G swap
/dev/dsk/c0t0d0s5 2.5G /tmp
Disk 1: 16 GB:
/dev/dsk/c0t1d0s0 7.7G /usr
/dev/dsk/c0t1d0s1 1.8G /opt
/dev/dsk/c0t1d0s3 3.2G /data1
/dev/dsk/c0t1d0s4 3.9G /data2
Disk 2: 33 GB:
/dev/dsk/c0t2d0s0 33G /data3
The data file systems are not in use right now, and I was thinking of
partitioning the data3 into 2 or 3 file systems and then creating
a new BE.
However, the system already has a BE (named s10) and that BE lists
all of the filesystems, incl the data ones.
# lufslist -n 's10'
boot environment name: s10
This boot environment is currently active.
This boot environment will be active on next system boot.
Filesystem fstype device size Mounted on Mount Options
/dev/dsk/c0t0d0s4 swap 4201703424 - -
/dev/dsk/c0t0d0s0 ufs 2098059264 / -
/dev/dsk/c0t1d0s0 ufs 8390375424 /usr -
/dev/dsk/c0t0d0s3 ufs 8390375424 /var -
/dev/dsk/c0t1d0s3 ufs 3505453056 /data1 -
/dev/dsk/c0t1d0s1 ufs 1997531136 /opt -
/dev/dsk/c0t1d0s4 ufs 4294785024 /data2 -
/dev/dsk/c0t2d0s0 ufs 36507484160 /data3 -
/dev/dsk/c0t0d0s5 ufs 2727290880 /tmp -
/dev/dsk/c0t0d0s1 ufs 770715648 /usr/openwin -
I browsed the Solaris 10 Installation Guide and the man pages
for the lu commands, but can not find how to remove the data
file systems from the BE.
How do I do a live upgrade on this system?
Thanks for your help.

Thanks for the tips.
I commented out the entries in /etc/vfstab, also had to remove the files /etc/lutab and /etc/lu/ICF.1
and then could create the Boot Environment from scratch.
I was also able to create another boot environment and copied into it,
but now I'm facing a different problem, error when trying to upgrade.
# lustatus
Boot Environment           Is       Active Active    Can    Copy
Name                       Complete Now    On Reboot Delete Status
s10                        yes      yes    yes       no     -
s10u6                      yes      no     no        yes    -        Now, I have the Solaris 10 Update 6 DVD image on another machine
which shares out the directory. I mounted it on this machine,
did a lofiadm and mounted that at /cdrom.
# ls -CF /cdrom /cdrom/boot /cdrom/platform
/cdrom:
Copyright                     boot/
JDS-THIRDPARTYLICENSEREADME   installer*
License/                      platform/
Solaris_10/
/cdrom/boot:
hsfs.bootblock   sparc.miniroot
/cdrom/platform:
sun4u/   sun4us/ sun4v/Now I did luupgrade and I get this error:
# luupgrade -u -n s10u6 -s /cdrom
ERROR: The media miniroot archive does not exist </cdrom/boot/x86.miniroot>.
ERROR: Cannot unmount miniroot at </cdrom/Solaris_10/Tools/Boot>.I find it strange that this sparc machine is complaining about x86.miniroot.
BTW, the machine on which the DVD image is happens to be x86 running Sol 10.
I thought that wouldn't matter, as it is just NFS sharing a directory which has a DVD image.
What am I doing wrong?
Thanks.

Kernel: PANIC! -- best practice for backup and recovery when modifying system?

I installed NVidia drivers on my OL6.6 system at home and something went bad with one of the libraries. On reboot, the kernel would panic and I couldn't get back into the system to fix anything. I ended up re-installing the OS to recovery my system.
What would be some best practices for backing up the system when making a change and then recovering if this happens again?
Would LVM snapshots be a good option? Can I recovery a snapshot from a rescue boot?
EX: File system snapshots with LVM | Ars Technica -- scroll down to the section discussing LVM.
Any pointers to documentation would be welcome as well. I'm just not sure what to do to revert the kernel or the system when installing something goes bad like this.
Thanks for your attention.

There is often a common misconception: A snapshot is not a backup. A snapshot and the original it was taken from initially share the same data blocks. LVM snapshot is a general purpose solution which can be used, for example, to quickly create a snapshot prior to a system upgrade, then if you are satisfied with the result, you would delete the snapshot.
The advantage of a snapshot is that it can be used for a live filesystem or volume while changes are written to the snapshot volume. Hence it's called "copy on write (COW), or copy on change if you want. This is necessary for system integrity to have a consistent data status of all data at a certain point in time and to allow changes happening, for example to perform a filesystem backup. A snapshot is no substitute for a disaster recovery in case you loose your storage media. A snapshot only takes seconds, and initially does not copy or backup any data, unless data changes. It is therefore important to delete the snapshot if no longer required, in order to prevent duplication of data and restore file system performance.
LVM was never a great thing under Linux and can cause serious I/O performance bottlenecks. If snapshot or COW technology suits your purpose, I suggest you look into Btrfs, which is a modern filesystem built into the latest Oracle UEK kernel. Btrfs employs the idea of subvolumes and is much more efficient that LVM because it can operate on files or directories while LVM is doing the whole logical volume.
Keep in mind however, you cannot use LVM or Btrfs with the boot partition, because the Grub boot loader, which loads the Linux kernel, cannot deal with LVM or BTRFS before loading the Linux kernel (catch22).
I think the following is an interesting and fun to read introduction explaining basic concepts:
http://events.linuxfoundation.org/sites/events/files/slides/Btrfs_1.pdf

Best Practice for setting bind variable when application loads

I am using JDeveloper 11.1.2.3.
When my application loads, the first unbounded page has a table populated by a named query.
I would like to set the parameter used by the named query when the page loads, to populate the initial data that is displayed.
What is the best practice for a solution to this issue?

user6003393 wrote:
I am using JDeveloper 11.1.2.3.
When my application loads, the first unbounded page has a table populated by a named query.
I would like to set the parameter used by the named query when the page loads, to populate the initial data that is displayed.
What is the best practice for a solution to this issue?Hi,
You can set the bind variable on VO by overriding prepareSession() method in Application Module check this http://docs.oracle.com/cd/E37975_01/web.111240/e16182/bcservices.htm#sthref357
Setting bind variable on runtime http://docs.oracle.com/cd/E37975_01/web.111240/e16182/bcquerying.htm#CHDECJHD
Zeeshan

Best Practices for FSCM Multiple systems scenario

Hi guys,
We have a scenario to implement FSCM credit, collections and dispute management solution for our landscape comprising the following:
a 4.6c system
a 4.7 system
an ECC 5 system
2 ECC6 systems
I have documented my design, but would like to double check and rob minds with colleagues regarding the following areas/questions.
Business partner replication and synchronization: what is the best practice for the initial replication of customers in each of the different systems to business partners in the FSCM system? (a) for the initial creation, and (b) for on-going synchronization of new customers and changes to existing customers?
Credit Management: what is the best practice for update of exposures from SD and FI-AR from each of the different systems? Should this be real-time for each transaction from SD and AR (synchronous) or periodic, say once a day? (assuming we can control this in the BADI)
Is there any particular point to note in dispute management?
Any other general note regarding this scenario?
Thanks in advance. Comments appreciated.

Hi,
I guess when you've the informations that the SAP can read and take some action, has to be asynchronous (from non-SAP to FSCM);
But when the credit analysis is done by non-SAP and like an 'Experian', SAP send the informations with invoices paid and not paid and this non-SAP group give a rate for this customer. All banks and big companies in the world does the same. And for this, you've the synchronous interface. This interface will updated the FSCM-CR (Credit), blocking or not the vendor, decreasing or increasing them limit amount to buy.
So, for these 1.000 sales orders, you'll have to think with PI in how to create an interface for this volume? What parameters SAP does has to check? There's an time interval to receive and send back? Will be a synchronous or asynchronous?
Contact your PI to help think in this information exchange.
Am I clear in your question?
JPA

Best Practice for Flat File Data Uploaded by Users

Hi,
I have the following scenario:
1.     Users would like to upload data from flat file and subsequently view their reports.
2.     SAP BW support team would not be involved in data upload process.
3.     Users would not go to RSA1 and use InfoPackages & DTPs. Hence, another mechanism for data upload is required.
4.     Users consists of two group, external and internal users. External users would not have access to SAP system. However, access via a portal is acceptable.
What are the best practice we should adopt for this scenario?
Thanks!

Hi,
I can share what we do in our project.
We get the files from the WEB to the Application Server in path which is for this process.The file placed in the server has a naming convention based on ur project,u can name it.Everyday the same name file is placed in the server with different data.The path in the infopackage is fixed to that location in the server.After this the process chain trigers and loads the data from that particular path which is fixed in the application server.After the load completes,a copy of file is taken as back up and deleted from that path.
So this happens everyday.
Rgds
SVU123
Edited by: svu123 on Mar 25, 2011 5:46 AM

Best Practices for image layout and positioning using anchored frames in Framemaker 10

Hi,
I'm looking for the best practices in how to layout my images in Framemaker 10 so that they translate correctly to Robohelp 9. I currently have images inside of Anchored frames that "Run into" the right side of my text. I've adjusted the size of the anchored frame so that my text flows correctly around the image. Everything looks good in Framemaker! Yeah! The problem is that when I link my Framemaker document to Robohelp, the text does not flow around my image in the same manner. On a couple of Robohelp screens the image is running into the footer. I'm wondering if I should be using tables in Framemaker in order to get the page layout that I'm looking for. Also, I went back and forth...is this a Framemaker question or is this a Robohelp question. Any assistance would be greatly appreciated.

I wish I could offer some advice, rather than simply adding to your question. I think there is something odd that happens with anchored frames.
I have been going through fits trying to figure out why the placement of my graphics shifts when I run them through Frame 10-to-RoboHelp 9 conversion. The placement in our books is flush left to the body text area. However, when they convert, some are flush left, some are centered, and a very few are flush right. My boss is very unhappy with me, as I have been unable to figure out why this is happening. I didn't create these files, so I don't know if it's something that goes back to how the graphics were initially imported. But I can't figure out why everything looks right in Framemaker, with the frame attached to an anchor tag, etc. but the placement goes hinky when they convert.
Any insights are appreciated. I'm wondering if it's going to come down to deleting them and recreating the graphic frame.

Best Practices for Exporting Files??

I'm new to Premiere (coming from FCP). I used Premiere months ago to compress some ProRes files to h.264 files for the web. I sent the files through Media Encoder and everything seemed fine. However, I realized after several weeks that the audio in all of the files was a few frames out of sync. Having not been a Premiere user at the time I did not do much research and decided to just use MPEG Streamclip from then on.
Now that I'm learning how to use Premiere, I looked up the issue on the forums and found that many people have had similar issues with the audio being out of sync after exporting. However, there are tons of different scenerios in which it seems to be occuring. The one common variable that I've noticed (among many of the threads, but not all) is that many of the people are exporting to a Quicktime format.
While I don't remember all the details of my export and sequence settings from my issue months ago (so I don't want to address that specific case), I am curious as to what are some "Best Practices" when exporting from Premiere Pro? Is there any advantage/disadvantage to use AME rather than exporting directly from Premiere Pro? In general, I will just be exporting as H.264 files for the web, MPEG-2 for DVD, and ProRes 422 for After Effects (or sometimes to bring into MPEG Streamclip).
I shoot almost entirely in AVCHD, and usually at 1080p 30fps. I'm running CS5 on a Macbook Pro 15" 2.0 Quad Core i7 8GB RAM.
While the question may seem broad, my main concern that I want to avoid is having the audio out of sync. But also I just want to know of any important details to keep in mind to prevent other issues.
Thanks,
Mike

> I'm running CS5...
What specific version? We're up to 5.0.4 now.
There have been bug fixes for audio/video synch in the updates. One of the fixes was for a bug in the conforming of audio and indexing of MPEG files, so you need to delete your media cache files and let Premiere Pro create new ones for this fix to take effect.

Best practice for setting an environment variable used during NW AS startup

We have installed some code which is running in both the ABAP and JAVA environment, and some functionality of this code is determined by the setting of operating system environment variables. We have therefore changed the .sapenv_<host>.csh and .sapenv_<host>.sh scripts found in the <sid>adm user home directory. This works, but we are wondering what happens when SAP is upgraded, and if these custom shell script changes to the .sh and .csh scripts will be overwritten during such an upgrade. Is there a better way to set environment variables so they can be used by the SAP server software when it has been started from <sid>adm user ?

Hi,
Thankyou. I was concerned that if I did that there might be a case where the .profile is not used, e.g. when a non-interactive process is started I was not sure if .profile is used.
What do you mean with non-interactive?
If you login to your machine as sidadm the profile is invoked using one of the files you meant. So when you start your Engine the Environment is property set. If another process is spawned or forked from a running process it inherits / uses the same Environment.
Also, on one of my servers I have a .profile a .login and also a .cshrc file. Do I need to update all of these ?
the .profile is used by bash and ksh
The .cshrc is used by csh and it is included via source on every Shell Startup if not invoked with the -f Flag
the .login is also used by csh and it is included via source from the .cshrc
So if you want to support all shells you should update the .profile (bash and ksh) and one of .cshrc or .login for csh or tcsh
In my /etc/passwd the <sid>adm user is configured with /bin/csh shell, so I think this means my .cshrc will be used and not the .profile ? Is this correct ?
Yes correct, as described above!
Hope this helps
Cheers

EP Upgrade - SP14 - Best Practice for Modification File Comparison

SDN Experts -
We are upgrading our EP from SP14 - SP16. SAP offers a file "diff" tool that is only useful for Java application files to assist in re-applying our mods on top of the new code stack.
We are looking for best practices in Portal upgrades to do the following:
- Identify all files that we have modified on existing SP
- Diff all source code files (java, XML, GUI, other) between Current SP14 and SP16
We are also looking for documentation that identifies the local directory structure for NWDS. This would aid us in creating a batch process to "diff" our source code libraries.
Any recommendations are appreciated.
Thanks

I'm not realy getting your question because you already state what to do:
We are looking for best practices in Portal upgrades to do the following:
Identify all files that we have modified on existing SP
Diff all source code files (java, XML, GUI, other) between Current SP14 and SP16
You should know by documentation what is changed I guess? Then start diff-ing the code and recompile or repackage. NWDS also has diff functionalities.
Good luck,
Benjamin

Best practice about dial-peer creating when using analog lines

Hi,
I am trying to find out what is the best practice when creating dial-peer for analog lines on CME, should I use trunk group or create separate dial-peer for each FXO ports? If I use trunk group, is there any advantage ( lesser dial-peer) or disadvantage?
Thanks!

The advantage of trunk groups is that a single dial peer can point to for instance PSTN, rather then multiple dialpeers, with varying preference, each pointing to a separate FXO. Funtionally I can't see much difference. So I guess it also comes down to personal preference.
=============================
Please remember to rate useful posts, by clicking on the stars below.
=============================

Best practices for audio files in your app

Hello All,
I was just curious what everyone is doing as far as having quality sounds in their app but keeping the file size down. Here are a few questions:
Is it better to have them compressed like mp3's, or not as .wav. Is .caf really better than .wav?
What have you found to be the best "sweet spot" for sample rate etc. for good quality and lower file size for an iphone app?
What programs can you recommended to do the final conversions necessary to keep the file size low and the file still working for the iphone. As an example when I make a .wav file by bouncing a mix from my PC based protools it creates .wav file that for some reason throws an error when used in an app; but if I take that same .wav file and open it in freeware like Audacity and save it as a .wav again; it reduces the file size and "fixes" whatever the issue was.
Any help you can offer would be very much appreciated. Thank you.

Well I'm not sure if any of the benefits of caf will be apparent to your app given that you are playing and not recording and you are using PCM. I think you can go directly from wav to caf with /Developer/Examples/CoreAudio/SimpleSDK/ConvertFile but I'm not sure. However, if you can produce compressed audio you will have a lot of space savings and you can experiment with sample rates to make a good size / quality tradeoff. I do not know what tools there are to record compressed audio...

Best practices for saving files when emailing to the media?

I'm pretty much an Ai imposter. I have Illustrator cs4 (14.0.0) and the way I use it is to cobble my designs together from iStock vector files and a lot of trial and error. Occasionally my husband throws me a lifeline because he is skilled with PhotoShop. As a small business owner, I am dying to find the time to take a course but right now I am just trying to float around the forums and learn what I can. Basically I flail about near the computer and the design sloooowly and somewhat mysteriously comes together. It's getting faster and is kind of fun, unless I get stuck.
This week I had a problem with an ad I designed for a very small community orchestra's printed programs. I send them a pdf first, and when they view it on their screens using inDesign they see the whole image but when it printed it omitted an element (a sunburst in the background). They thought it had something to do with that element being in color rather than greyscale (though there were other elements that survived and they were the exact same color so I was skeptical). I sent a greyscale file, no luck. I sent them the Ai file, but that apparently "crashed" their iD and now they believe I've sent a corrupted file. They aren't very adobe-savvy, either.
I've designed and emailed no fewer than 9 other ads to other print & online media organizations this year and never had a problem. The file looks fine to me in all versions I open/upload/email to myself.
You can see it here if you like: http://www.scribd.com/doc/27776704/RCMA-for-CSO-Greyscale
So here are my specific questions:
1. How SHOULD I be saving this stuff? Is pdf the mark of a rookie?
2. What settings should I be looking at when/before saving? I read about overprint, for example, and did try that with one of the versions I sent them to no avail. I don't really know what that does, so I was just trying a hail mary there anyway.
Thanks for your time!

PDF is the modern way of sending files but what you might want to do is in this case select the art in question and go to Object>Flatten Transparency
then save it as a pdf or as an ai file when sending the file to them zip it if they have windows based computers or stuffit if they have Mac actually zip is good for both.
It is safer to send it as an archive then as an ai file.

Best Practice For Secure File Sharing?

I'm a newbie to both OX X Server and File Sharing protocols, so please excuse my ignorance...
My client would like to share folders in the most secure way possible; I was considering that what might be the best way would be for them to VPN into the server and then view the files through the VPN tunnel; my only issue with this is that I have no idea how to open up File Sharing to ONLY allow users who are connecting from the VPN (i.e. from inside of the internal network)... I don't see any options in Server Admin to restrict users in that way....
I'm not afraid of the command line, FYI, I just don't know if this is:
1. Possible!
And 2. The best way to ensure secure AND encrypted file sharing via the server...
Thanks for any suggestions!

my only issue with this is that I have no idea how to open up File Sharing to ONLY allow users who are connecting from the VPN
Simple - don't expose your server to the outside world.
As long as you're running on a NAT network behind some firewall or router that's filtering traffic, no external traffic can get to your server unless you setup port forwarding - this is the method used to run, say, a public web server where you tell the router/firewall to allow incoming traffic on port 80 to get to your server.
If you don't setup any port forwarding, no external traffic can get in.
There are additional steps you can take - such as running the software firewall built into Mac OS X to tell it to only accept network connections from the local network, but that's not necessary in most cases.
And 2. The best way to ensure secure AND encrypted file sharing via the server...
VPN should take care of most of your concerns - at least as far as the file server is concerned. I'd be more worried about what happens to the files once they leave the network - for example have you ensured that the remote user's local system is sufficiently secured so that no one can get the documents off his machine once they're downloaded?

What are best practices for back up drives when they become full?

I've got an external hard drive that is now full. What do I do? Delete the oldest files on it, right? If so, do I delete the full backup and ALL the incremental backups since then? Do I have to do this manually now; isn't there a way to automate this?

The "usual" procedure is to move the oldest incremental backups from the hard drive to permanent storage media such as DVD.
However, if your backup drive is the same size as your main drive, then you could do incremental backups that delete files on the backup druve that have been deleted from the source drive. This prevents the drive from filling up with old and out of date files. However, if you need to maintain archives of even the incremental backups, then move the oldest to permanent storage media and delete them from the backup drive to free space.
Visit The XLab FAQs and read the FAQ on backup and restore for more information.
Why reward points?(Quoted from Discussions Terms of Use.)
The reward system helps to increase community participation. When a community member gives you (or another member) a reward for providing helpful advice or a solution to their question, your accumulated points will increase your status level within the community.
Members may reward you with 5 points if they deem that your reply is helpful and 10 points if you post a solution to their issue. Likewise, when you mark a reply as Helpful or Solved in your own created topic, you will be awarding the respondent with the same point values.

Best practices for ZFS file systems when using live upgrade?

Similar Messages

Maybe you are looking for