Qmaster and XSAN?

I'm trying to set up Final Cut Server on our fibre server and in the setup and installation menu it mentions creating quickclusters and custom clusters. Are these settings meant only for the server computer or all computers connected to the network? What are some best practices with this? If I am sending bits to other computers directly I'm definitely going over giga-ethernet, but if I'm going fibre, wouldn't I have to copy once to the server computer, then again to the cluster computers?

You are correct. I only do the Gig-E configuration when customer don't care about how long the comnpression takes.
But normally all my deployment are done over Xsan.
So the tips for Xsan are as follows:
On the cluster controller, point the cluster storage to a folder on the Xsan volume, make sure that all the other nodes users and group account in the cluster have full read/write access to this location. On the other nodes configure there cluster storage to that directory as well. This way no copying of source media is ever made to the cluster node.
Does this make sense?

Similar Messages

  • File Permission problems FCP and XSAN

    We have 5 Mac Pros connected to our XSAN (2 Xserves and 2 Promise arrays.). When a file is created on one of the FCP Macs and saved on the XSAN, the permission are OK (Individual - R/W, Group - R/W & Everyone RO.)
    The problem is when a file, generally audio (from our sound booth or music from an outside source) or a graphic (usually Photoshop from our Graphics folks) is brought in over the LAN and save to the XSAN volume, the permissions are Individual - R/W, no group is listed and Everyone - No Access.
    If they try work on something that another user started, they get permission errors if the sequence contains a file that was brought in from the outside.
    I know we can change permissions on a file, but there are so many of them that it is slowing down production.
    Is there anything we can do to either eliminate this happening, or have something like a drop folder they can place a file in to automatically change the file permissions?
    Any help with this would be greatly appreciated.
    The MacPros are running OS X 10.5.8 and FCP 7.0.3 and the XSAN is running Mac OS X Server 10.5.8 and XSAN 2.2.
    Thanks.

    If files saved to the AFP reshare are inaccessible by the fibre clients try this in terminal on your AFP re-share...
    sudo serveradmin settings afp:lock_manager = no
    Then restart AFP (or reboot), and re-propagate permissions back to where you need them. This will hand off file locking duties from AFP to the Xsan, which will help keep your permissions correct.
    JM
    (Thanks to Shane for showing me this)

  • Qmaster and AE6.5

    Hi,
    I have created already a cluster controller with 11 service nodes, installed already the AE6.5, the necessary plugins and qmaster software on all nodes, created also a cluster in qadministration.
    When i tried to submit batches in my cluster it is successful but the output files is no where to be found. can anyone help me with this qmaster and AE.
    I have read also the forum here about qmaster and AE done everything they told, mounted all the shred drive but still no output files. though my cluster is running, receiving jobs and processing and generating an status of Succesful Progress... After Effects succesfully launched
    this is my unix template
    /Applications/Adobe\ After\ Effects\ 6.5/aerender -project /Volumes/render-1.local/project/test6.5\ folder/test6.5.aep -comp "Comp 1" -RStemplate "Multi-Machine Settings" -OMtemplate "Multi-Machine Sequence" -s 1 -e 25 -i 1 -output /Volumes/render-1.local/output/test[#####].psd
    help anybody!

    hi johnaris,
    in your unix template:
    /Applications/Adobe\ After\ Effects\ 6.5/aerender -project /Volumes/render-1.local/project/test6.5\ folder/test6.5.aep -comp "Comp 1" -RStemplate "Multi-Machine Settings" -OMtemplate "Multi-Machine Sequence" -s 1 -e 25 -i 1 -output /Volumes/render-1.local/output/test[#####].psd
    there is basically nothing wrong with it. your render nodes are working well. the reason why only one service node is rendering your submitted job is that, you were only submitting 25-frames to render.
    By default, qmaster allocates 10-frames each segment (minimum number of frames in each apple qmaster segment). Because your submitted job is 25-frames, Qmaster will divide them into 2 segments of 10 frames each. Mac Minis have 2 cpus, therefore the frames were being accomodated into one mac mini.
    For this, try lowering the number of minimum value to 5 under the "Minimum" field. This way, apple qmaster divides the job into 5-segments of 5-frames each. You should have 3 service nodes running with this scenario.
    If you like, try setting the minimum value to 1 so qmaster divides the job into 25-segments of 1-frame each. With this, all your 11-macminis (service nodes) should be working.
    regarding that "unix error command 9" -- no idea...
    hope that helps.
    cheers!
    nante

  • Lost IP address on G5 w/ OS X server 1.4 and XSAN 1.3

    I attempted to change the DHCP setting via Chicken of the VNC on a G5 loaded with OS X 1.4 and Xsan 1.3 and now am unable to determine what the IP address is. I had my network manager revert the port to a set IP range and sniffed that with Ethereal with no luck. I can start the G5 with firm ware control so that is shows up via firewire on my ibook laptop as two server HDs but is there a way to determine its IP from searching the server HDs or is there another way to get full access to the server via firewire. I guess what would be ideal is if there is a way to fully manage the G5 over firewire. I am pretty new at this so any help would be appreciated even if it seems painfully obvious to you all.
    Call Center   Mac OS X (10.4.6)   4 XSAN G5s with 8 XRAID Arrays

    connected via firewire and re-configured the preferences plist to correctly pull the IP via IP allowing me to collect the IP useing ethereal.

  • Can't mount 3rd party DAS storage and Xsan volume simutaneously

    Xsan client(MacPro) equipped an Apple FC HBA card can't mount SGI TP9300 DAS storage and Xsan volume simutaneously when powering on or restarting.
    Among 2 FC ports, one port is used for Xsan volume using Xserve RAID, and the other port is used for SGI TP9300 storage which is directly connected to that Mac.
    When booting, Xsan volume is always mounted normally, but SGI TP9300 volume is not.
    Sometimes TP9300 volume is mounted about 20~30 minutes later.(I attached system.log file)
    The LUN of TP9300 is seen at Disk Utility, so it's not physical problem.
    It seems that actual volume information is not correctly processed at booting time.
    While, after booting the system with the FC cable for TP9300 disconnected and confirming that Xsan volume mounted normally, then connecting the FC cable for TP9300, the TP9300 volume is mounted immediately.
    Mac OS X 10.4.8, Xsan 1.4 is being used.

    Yes, I think you screwed your xsan volume. You can not remove luns from pools, and you can not remove volumes from volumes.
    How important is your existing data? In your case I would concentrate on that. This is not trivial. Theorotically you have lost your data allready. In practise you might have luck.
    Once I was in your kind of twilight zone, and managed to keep the data. It is a complete cli job. You have to be very carefull about with your config files. You could hire an experienced xsan consultant. If that is not possible, read the manpages very carefully before doing anything on the volume.
    Regards
    Donald

  • How to upgrade to 10.4 Tiger and Xsan 1.3

    I am running Xsan 1.1 on and Xserve G5/RAID running OS X 10.3.9. I need to get everything up to date so that client machines can be running 10.4.x Tiger. What is the best upgrade path? ie...
    - do I upgrade server to Tiger first, still running Xsan 1.1 (for 10.3.9)?
    - do I upgrade to Xsan 1.3 on 10.3.9 first, and then upgrade to 10.4?
    Also.. I still have to upgrade the RAID firmware to 1.5 because after downloading and installing RAID 1.5, it won't launch. Anybody have ideas on that? It just bounces in the doc twice and then quits. So.. I can't update the firmware.

    Thanks for everyone's guidance on this process. I completed my upgrade yesterday. All is running smoothly. I thought I would post a brief description of my experience which may help others.
    Background:
    I have a very basic Xsan setup: 1ea Xserve G5 controller, 2ea G5/Dual 2.5GHz clients, a Sanbox fibre switch, and 2ea Linksys gigabit switches. All three machines were running Xsan 1.1 on OS X 10.3.9 Panther. I needed to get one of my two client machines running Tiger. So... due to Xsan compatibility issues, I had to upgrade everything to OS X 10.4.6 Tiger and Xsan 1.3. I also needed to upgrade my Xserve RAID to the latest firmware (v1.5).
    Upgrade Process:
    I began this process with all three machines running. Although the Xsan volume was 'running', it was not mounted on any of the machines. My Xserve is designated as the primary controller (High priority); both clients are backup controllers (Med and Low priority).
    Here's what I did to get to 10.4.6 and Xsan 1.3:
    1) Upgraded the XServe to OS X Server 10.4 using install DVD. This required a server restart which caused Xsan to switch to my 'Med' client machine as the controller. I verified this by running Xsan Admin 1.1 on the client machine; it saw only the two clients (just as it should).
    2) After restart, used Software Update on the Xserve to bring it up to OS X Server 10.4.6. This included the 'Xsan 1.2 Update for 10.4' and 'Xsan Admin 1.2 Update' along with a bunch of other misc updates (Java, WebObjects, Security..). Restarted.
    3) Ran Xsan Admin 1.2 on the server. It was controlling the Xsan volume (which was still running). Both clients were still visible as backup controllers. I did not attempt to mount the Xsan volume
    4) Installed OS X 10.4.3 Tiger Upgrade from DVD on Client1. Restarted.
    5) After restart, used Software Update to bring Client1 up to OS X 10.4.6. This included the 'Xsan 1.2 Update for 10.4' and 'Xsan Admin 1.2 Update' along with a bunch of other misc updates (Java, Garageband, QuickTime). Restarted.
    6) Used Software Update to bring Client2 up to OS X 10.4.6. This included the 'Xsan 1.2 Update for 10.4' and 'Xsan Admin 1.2 Update' along with a bunch of other misc updates (Java, Garageband, QuickTime). Restarted.
    7) So.. at this point everything is running OS X 10.4.6 (the latest OS) and Xsan 1.2. And.. Xsan Admin 1.2 still running on the server. But... for some unknown reason, the 'Clients' tab no longer shows the Server as an available client. It only shows my two G5's as clients. Wierd and scary!! I quit and relaunched Xsan Admin 1.2, did a 'Refresh, but it made no difference. The '1. Computers' tab for the SAN itself does show all three computers (as controllers), but the 'Clients' list for this volume only shows my two client workstations (both 'Not mounted'). I decided to hold my nose and continue (such is life with Xsan...).
    8) Used Xsan Admin 1.2 on the Server to 'Stop' the Xsan volume.
    9) Launched Terminal on the Server and ran the 'cvfsck' disk utility on my Xsan volume. Here is what I typed into terminal:
    sudo cvfsck Napoleon
    (my volume is named 'Napoleon') I did have to authenticate with an Admin-level password. The utility took 5-10 min to run on my 5TB raid. Most of the time is just showed a static cursor in the Terminal window, but then the errors appeared. It reported 50+ errors, all of the same type. Here is a sample of what each error was like:
    Error: Inode 0x804caa865e6e nsubnirs count is 34 - should be 18. Fixed.
    The utility completed with all errors 'fixed.' Quit out of Terminal.
    10) Used Xsan Admin 1.2 on the Server to 'Start' the Xsan volume back up.
    11) Used Software Update on the Xserve to install the 'Xsan 1.3 Update for 10.4' and 'Xsan Admin 1.3 Update'. Restarted.
    12) Ran Xsan Admin 1.3 on the Server. Now.. the Server does show in both the Computers list (as the primary controller) and in the Clients list (still not mounted, of course).
    13) On Client1, used Software Update to install the 'Xsan 1.3 Update for 10.4' and 'Xsan Admin 1.3 Update' along with a bunch of other misc updates (iCal, iDVD, etc). Restarted.
    14) On Client2, used Software Update to install the 'Xsan 1.3 Update for 10.4' and 'Xsan Admin 1.3 Update' along with a bunch of other misc updates (iCal, iDVD, etc). Restarted.
    15) So.. now everything is running OS X 10.4.6 Tiger and Xsan 1.3. Xsan Admin 1.3 on the Server shows all three machines as both Controllers and Clients. The Xsan volume is 'running' but not mounted on any of the machines.
    16) Used Xsan Admin 1.3 on the Server to 'Stop' the Xsan volume.
    17) Launched Terminal on the Server and ran the 'cvfsck' disk utility on the Xsan volume. Here is what I typed into terminal:
    sudo cvfsck Napoleon
    This time no errors were reported, although terminal did produce a very nice report detailing what was going on as the Xsan volume was checked (thank you Apple!). Quit Terminal.
    18) Used Xsan Admin 1.3 on the Server to 'Start' the Xsan volume back up.
    19) And now for the big moment... Used Xsan Admin 1.3 on the Server to mount the Xsan volume on Client1. It took a little longer than nomal, but it did mount. I immediately browsed through the volume and copied a large folder to the Client1 desktop. Worked great. Copied the same folder back to the Xsan volume. Worked great. Deleted the folder in both places and emptied Trash. Worked great. I repeated this same mount/copy/delete process with Client2. Everything was working great!!
    Now it was time to upgrade the RAID firmware to v1.5:
    20) Used Xsan Admin 1.3 to unmount the Xsan volume on both clients. I then 'stopped' the volume and quit out of Xsan Admin.
    21) Downloaded the RAID Admin 1.5 update and installed it.
    22) Ran RAID Admin 1.5 and connected to my Xserve RAID. Used the menu command 'Update firmware...' to begin the process. It asks you to locate the firmware source file on disk. The entire update process took about 15 min to complete. The RAID restarted after the update. After seeing all green lights on the RAID, I quit RAID Admin 1.5.
    23) Launched Xsan Admin 1.3 on the Server to 'Start' the Xsan volume.
    24) Used Xsan Admin 1.3 on Server to mount the Xsan volume on both clients. Both mounted just fine.
    And now I'm up to date
    I know this is a long description, but I wish that I had had someone else's detailed log before I headed into this overly complex upgrade process (Apple... please make this easier going forward!). Hope it helps someone else.
    John Christensen
    [email protected]
    817.690.8385
    G5   Mac OS X (10.3.9)   XServe, XServe Raid, Xsan

  • FCP and Xsan

    Is anyone working with FCP and XSAN?
    I working on building a new Server, XSAN system and I am curious on how your workflow is working.

    Our workflow is working fine. This discussion is best left for the XSAN forum.

  • Apple QMaster and QAdministrator

    Hi,
    Does anyone here use the Apple QMaster and QAdministrator software to allow rendering accross multiple MAC's? Does this even work for After Effects? I have three MAC's here at the office, (slightly different specs accross the board) and I was wondering if there is a way to set them up to work as a sort if render farm? That way, I can create a few projects on one, and schedule some rendering to happen, and make the most use of the idle machines.
    Can someone point me in the right direction on how to go about creating a render farm, what is involved, and where to begin?
    My previous posts about the hardware I wanted to get, and the replies are making me a little concerned that I am going to waste a large amount of money on machines we dont need yet. (at this point, one of my biggest concerns is that we need a license for Twixtor, but can either get MAC, or Windows, they will not allow us to intechange if we decide to go the PC route. So I need to be very sure of my facts before I send the moneu on the Twixtor license)
    Thank you
    Pierre

    Hi,
    Thanks again! - I have sent some mails to Re:Vision contact centre to weigh up our options. We are still working on the first part of the project, and being self-funded, our budget is a little tight. Honestly, when we go live, and we believe it will be successful.. I would like to get as many plug ins as my little heart desires. I would love all of Video CoPilots stuff, and Red Giant, and Boris, and+and+and!
    For now, I am being a little cautious though. So, we need twixtor to make up the extra slow motion our camera cant do, but other than that, as far as After Effects special effects go, we will probably try to use mostly stock footage for the first part. Our greatest "Issue" is going to be rendering time - I will look into your suggestions, as see what we can come up with. I am going to hand over my to IMAC's and get myself two powerhouse PC's instead. The MACS will go back to the the other guys, so we will still use them, just not for now. So all licenses will be PC licenses to start with - untill we make the big time, and hire the entire forum to come work with us!
    Take it easy,
    and thanks agan for the advice.
    Pierre

  • What is Qmaster and Qadministrator?

    I need some advice about two applications that suddenly appeared in my application folder. Qmaster and Qadministrator. I see that they are Apple programs, but don't know how they got there (though they may have come as part of some other programs I recently installed) nor do I know how to use them or if I really need to use them. I see that there is a referrence to networking, but I am not on a network per se (other than a router connecting my desk top Mac to my Macbook.
    I searched the forums and can't seem to find what this does, how it got there, and how it works.
    Thanks.

    They are components of Final Cut.

  • 10.6.7 and XSAN 2.2.1

    Hi all,
    Is there anybody able to run server 10.6.7 and Xsan 2.2.1?

    Not yet. We are still on 10.6.4 as each new update has had some sort of significant issue. These things don't directly effect the function of an MDC, but since your clients have to run an equivalent version of OS X or older, we haven't recommended that any of our clients update their MDCs to accomodate an OS update on their client machines.
    10.6.5, Wifi:  https://discussions.apple.com/thread/2644274?start=0&tstart=0
    10.6.6, FCP/ ProKit6: http://www.digitalrebellion.com/blog/posts/mac_os_x_10.6.6_breaking_final_cut_st udio.html
    10.6.7, Fonts in FCP: https://discussions.apple.com/thread/2792142?start=30&tstart=0
    JM

  • Qmaster and Foundry Nuke v5 - solution

    For those who is interested in integrating The Foundry Nuke into Qmaster pipeline here is the solution.
    It's really easy:
    1. Instal applications, bla bla bla (as usual)
    2. Open Qmaster and select Generic render
    3. Create new template and add the following commands executable,input, start frame, end frame, steps (in Format panel)
    4. in Executable panel ( on top ) put "/Applications/Nuke5.1v2/Nuke5.1v2.app/Contents/MacOS/Nuke5.1v2" (without quotes, Nuke5.1v2 should be changed to the name YOU HAVE on your system)
    5. Input field - is a path to the Nuke project file (with .nk) you want to render
    6. Item field - empty
    7. Star and end fields - is the same as in project.
    8. Steps field - is a way to control which frame ( if you project is a sequences) will be rendered ( say you put 2 thats mean every second frame is rendered and so on)
    9. Minimum field - is a field what control how many frame will be given to each machine in a cluster.
    10. Now in Format field after the executable put -x ( that makes Nuke render in command line without opening the Nuke app)
    Now Render.....
    Good luck.
    Please post comments if you have any experiences with Nuke and Qmaster.
    Message was edited by: vadim007

  • Qualified success with FCS, qmaster and compressor, but...

    I finally have FCS, qmaster and compressor working "correctly". So long as I don't change the working directory from /var/spool, I can use either a managed-cluster or a quick-cluster. Thankfully my system disk is 300GB so I have room to play.
    The big issue I face is that qmaster does not seem to realize that my sources are "edit in place" and shared via sanMP. There is no reason to copy the originals to the servers spool in order to share them, but it does. I assume this could be changed in compressor, but that hasn't helped yet. It doesn't make a copy if the cluster is only on a single machine, but if I use services from another machine, it goes to town copying files. These are all really big prores captures, so this is not a desired behavior, even on a 10GB iSCSI SAN.
    Anyone have a clue?? Maybe I need to mess with compressor prefs some more?
    Thanks,
    Eric.

    I finally have FCS, qmaster and compressor working "correctly". So long as I don't change the working directory from /var/spool, I can use either a managed-cluster or a quick-cluster. Thankfully my system disk is 300GB so I have room to play.
    The big issue I face is that qmaster does not seem to realize that my sources are "edit in place" and shared via sanMP. There is no reason to copy the originals to the servers spool in order to share them, but it does. I assume this could be changed in compressor, but that hasn't helped yet. It doesn't make a copy if the cluster is only on a single machine, but if I use services from another machine, it goes to town copying files. These are all really big prores captures, so this is not a desired behavior, even on a 10GB iSCSI SAN.
    Anyone have a clue?? Maybe I need to mess with compressor prefs some more?
    Thanks,
    Eric.

  • What is QMaster and QAdmin?

    What exactly is Qmaster and QAdministrator? There is no explanation in Help or on the Leopard page. What does it do and how can I use it?

    http://en.wikipedia.org/wiki/Apple_Qmaster

  • Saving Files on FCP 6.0.2 and XSan

    Hello All,
    We're running into very frustrating issues as we are trying to save and open Final Cut Pro project files. We are currently using an XSan with ACLs set up to store FCP documents and related files. The six workstations aren't directly attached to the SAN, they connect via GigE to AFP shares (not an ideal solution, but it's what we can afford right now). It should be noted that our setup is currently using the "Golden Triangle" of directory services (AD user accounts and groups).
    In December we upgraded our XSan to version 2 (which was a nightmare in itself), and all of our clients were upgraded to 10.5 with FCS2. After the upgrades, a few issues have arisen:
    1. When a user attempts to save a FCP project, sometimes they receive the error message that the file is locked. From what I read on apples support site, this is due to permissions-- a file cannot be saved by one person, and then opened and saved by another. If the ACLs set on the folder reflect the groups that have full access to the files in which the project resides (recursively), why is this an issue? Or is this just another case of bad execution on apples part?
    2. After concluding a "Save As", most of the time the alternate project file is able to be opened. We've run into several instances, however, when the project file is a very small size (16k or so), and gives the error message "Unable to open project file".. and it's pretty much trash. In these cases, the autosaves are behaving in a similar manner, being a very low size and not opening.
    So my question is, does this all sound like permissions issues.. or is there something inherently wrong with the XSan at this point (e.g. corruption)? How are other people using XSan volumes in conjunction with FCP 6.0.2/FCS2 to edit and store files? Perhaps we're going about the whole setup the wrong way..
    Thanks for your time and any advice you can provide..

    Hi,
    Thanks for taking the time to reply.
    We have six producers.
    Each producer has their own storage folder on the SAN: Producer 1, Producer 2, Producer 3, Producer 4, Producer 5, Producer 6. Each of these folders are located in One sharepoint called "Producers".
    So we have one main producer, Well call Joan Smith. Joan has full access, recursively throughout her folder, which is producer 1. Joan has interns that work for her, who use an account called pro1edit1. Pro1edit1 has full access to Producer 1's folders as well, recursively. Joan and her intern never work on the same project at the same time from two different workstations. Each project is edited by one and only one user from one workstation.
    Either Joan or her intern logs into the Mac (using either Joan's username/password, or the intern using their generic username and password [pro1edit1]). These accounts are AD Accounts. After a login, OD sends MCX policies to the clients (such as auto-mounting the AFP share points ["Producers", for instance] with their username and password). The user browses to their folder. If they create a new project, the create a folder with the project name. They then open FCP and set the settings (scratch, etc) to point to that folder. They save a project file with the name of the project within this same folder, keeping everything as one general "package" in the folder. For instance: /Volumes/Producers/Producer\ 1/5-Minute-Piece/FiveMinutePiece.fcp. All of the other folders are also located in here, scratch disks, autosaves, captures, etc.
    When the user is done, they save, close final cut, and quit.
    The next say, either Joan or her intern come in to do some work on the project. They log into one of the workstations, and they open the file. They conclude some edits, and upon trying to save the file they receive the error message that the file is locked. They then save the file as an alternate filename. Sometimes this works. Sometimes it says it saves it fine, but the next attempt at opening the file yields the error that the project cannot be opened. Upon inspection of the file, it's about 16k in size.
    Using Producers folder/sharepoint, the permissions are as follows:
    Producers:
    <AD Domain>\grpStaffProduction: Allow: Read Only: This folder. *
    <AD Domain\grpProEdit: Allow: Read Only: This folder. **
    <AD Domain\grpProduction: Allow: Read Only: This folder. *
    <AD Domain\grpEditAdmins: Allow: Full Control: This Folder, Children, all descendants. [inherited]
    <AD Domain\grpDomainAdmins: Allow: Full Control: This Folder, Children, all descendants. [inherited]
    Local Administrator: Allow: Full Control: This Folder, Children, all descendants. [inherited]
    POSIX:
    admin: RW: This Folder
    staff: R: This folder
    Others: None: This folder
    Using Producer 1 (within the above folder), the permissions are as follows:
    <AD Domain>\JoanSmith: Allow: Full control: This Folder, Children, all descendants.
    <AD Domain>\pro1edit1: Allow: Full Control: This Folder, Children, all descendants.
    <AD Domain\grpEditAdmins: Allow: Full Control: This Folder, Children, all descendants. [inherited]
    <AD Domain\grpDomainAdmins: Allow: Full Control: This Folder, Children, all descendants. [inherited]
    Local Administrator: Allow: Full Control: This Folder, Children, all descendants. [inherited]
    POSIX:
    admin: RW: This folder.
    staff: R: This folder.
    Others: R: This folder.
    +* This group is used for another producer folder. This allows AFP to mount the share, and for the users to see their producer folder.+
    +** This group is used for another producer folder. This allows AFP to mount the share, and for the users to see their producer folder.+
    +* Joan, all other users in production, and their interns are apart of this group. This allows them to mount the AFP share and see which folders they have access to.+
    So that's some more detail on our permissions and general workflow. We're up for suggestions. I appreciate your response!
    -M

  • Leopard - QMaster and Virtual Cluster problem

    Hi guys,
    Up until yesterday I had my MacPro Octo running under 10.4 where I did succesfully set up a Virtual Cluster using 4 instances for Compressor. It worked as a charm and my MacPro was doing it's job perfectly.
    Today, I made a bootable backup of my 10.4 install and installed 10.5 using the erase and install options ( clean install ). I installed all my software again and tryed setting up my Virtual Cluster again, using the same settings I had under 10.4. Sadly I can't seem to get it working.
    In the QMaster Preferences pane, I have the QuickCluster with Services option checked. For the compressor entry in the Services I have the Share Option checked and used 4 isntances for the selected service. The Quickcluster received a descent name and the option to include unmanaged services from other computers is checked.
    I have the default options set in the Advanced tab, ( nothing checked except log service activity to log file and Show QMaster service status in the Menu Bar). I then started the Cluster using the Start Sharing button.
    Now I open u Compressor and add a file to process ( QT encode to iPod ), but when I hit the Submit button, my Virtual Cluster doesn't show up in the Cluster Dropdown. If I now leave the Compressor GUI open for 5 minutes, it will eventually show up in the list, and I can pick it. Sadly, picking it from the list at this point and hitting the Submit button makes Compressor Hang.
    I checked my logs, but the only thing concerning Compressor I could find is this :
    4/12/07 20:12:35 Compressor[242] Could not find image named 'MPEG1-Output'.
    4/12/07 20:12:35 Compressor[242] Could not find image named 'MPEG1-Output'.
    4/12/07 20:12:35 Compressor[242] Could not find image named 'MPEG1-Output'.
    4/12/07 20:12:41 Batch Monitor[190] * CDOClient::connect2: CException [NSException raised by 'NSPortTimeoutException', reason = '[NSPortCoder sendBeforeTime:sendReplyPort:] timed out (218488391.647220 218488361.647369) 1'], server [tcp://10.0.1.199:49167]
    4/12/07 20:12:41 Batch Monitor[190] exception caught in -[ClusterStatus getNewStatusFromController:withOptions:withQueryList:]: NSException raised by 'NSPortTimeoutException', reason = '[NSPortCoder sendBeforeTime:sendReplyPort:] timed out (218488391.647220 218488361.647369) 1'
    4/12/07 20:17:55 Batch Monitor[190] * CDOClient::connect2: CException [NSException raised by 'NSPortTimeoutException', reason = '[NSPortCoder sendBeforeTime:sendReplyPort:] timed out (218488705.075513 218488675.075652) 1'], server [tcp://10.0.1.199:49167]
    I tried Stopping and then Restart Sharing and I noticed the follwoing entries in my log :
    4/12/07 20:23:26 compressord[210] can't refresh cache from file "/Library/Application Support/Apple Qmaster/qmasterservices.plist"
    4/12/07 20:23:26 compressord[211] can't refresh cache from file "/Library/Application Support/Apple Qmaster/qmasterservices.plist"
    4/12/07 20:23:26 compressord[213] can't refresh cache from file "/Library/Application Support/Apple Qmaster/qmasterservices.plist"
    4/12/07 20:23:26 qmasterca[269] can't refresh cache from file "/Library/Application Support/Apple Qmaster/qmasterservices.plist"
    4/12/07 20:23:26 qmasterqd[199] can't refresh cache from file "/Library/Application Support/Apple Qmaster/qmasterservices.plist"
    4/12/07 20:23:27 QmasterStatusMenu[178] * CDOClient::connect2: CException [NSException raised by 'NSPortTimeoutException', reason = '[NSPortCoder sendBeforeTime:sendReplyPort:] timed out (218489009.603992 218489007.604126) 1'], server [tcp://10.0.1.199:49407]
    4/12/07 20:23:27 Batch Monitor[190] * CDOClient::connect2: CException [NSException raised by 'NSPortTimeoutException', reason = '[NSPortCoder sendBeforeTime:sendReplyPort:] timed out (218489037.738080 218489007.738169) 1'], server [tcp://10.0.1.199:49407]
    4/12/07 20:23:27 Batch Monitor[190] exception caught in -[ClusterStatus getNewStatusFromController:withOptions:withQueryList:]: NSException raised by 'NSPortTimeoutException', reason = '[NSPortCoder sendBeforeTime:sendReplyPort:] timed out (218489037.738080 218489007.738169) 1'
    Batch Monitor immediately detects the cluster being active again, but Compressor doesnt, leaving me only This Computer available in the Cluster drop down when submitting a batch.
    In my Activity Monitor, I notice that CompressorTranscoder is not responing ( the 4 CompressorTranscoderX processes are fine ) and the ContentAgent proces isn't responding neither.
    Does anyone have any clue on what I could check next or how I could fix my problems ?
    Thanks a lot in advance,
    Stefaan

    Bah, this is crazy, today it doesn't work anymore. Yesterday my cluster was showing up in the Dropdown window, and I could submit a batch to it, and it got processed over my virtual cluster.
    Today, after finishing the second part of my movie, I tried it again. I didn't change anything to my settings, my machine hasn't even rebooted (just recovered from sleep mode) and my cluster isn't showing up at all anymore. Even the Qmaster menu doesn't show it
    Guess, I'll have to wait out until it appears again, or try a few things out

Maybe you are looking for