Re: [SunRay-Users] What are sunray users going to do once Oracle stop support
Clift, Tom CIV NSWCDD, K55 writes: I thought it was only software also.we have a large number of Sun Ray 3's that we will need to be replacing. I almost just need an appliance that will only run a browser and point the users to a SSGD server. Any thoughts? We used SGD. I think you will be much happier with thinlinc. We run from Linux-based HP thin clients (e.g., t610). -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] What are sunray users going to do once Oracle stop support
Clift, Tom CIV NSWCDD, K55 writes: So your thin clients connect to a thinlinc server? I am researching thinlinc know as I don't know much about it. Yes. The Linux-based HP t610 works well with ThinLinc, and we also chose these because we connect to VMware View. I had responded to a post about Sun SGD, which we used. There might be a scenario where SGD wins, but the X server in thinlinc works with more applications of interest to us (e.g., Maple) than that in SGD or the Ray. There is a web (html5) interface to thinlinc, so thinlinc is superficially similar to SGD. The thinlinc client has more features and is required for demanding audio/video (e.g., movie trailers). -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] USB device successfully mounted but not notified
Nicolás nico...@devels.es writes: Is there a way to configure automatic notifications for the users mounting their devices on their sessions? Is that a job for the utmountd daemon or is there something it's not working as intended? We've used # usbdrived # # Author: Daniel Cifuentes # Version: v67, July 2008 -- --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] { SPAM 1 }:{ SPAM 1 }: Success obtaining Oracle Premier Support for Software on small Sun Ray deployments?
Jayme Snyder writes: My requirements are simply for a multi-monitor thin client environment which can play a Youtube clip without major lag on Windows RDP or a VMware view. My opinion is that you will spend substantial time getting something unlikely to meet your requirements. Oracle has abandoned the Ray. We run HP t610 thinclient (Linux-based) with VMware View, along with a test of Cendio's thinlinc as a Ray replacement for connections to Windows and Unix systems. We love the Rays (on Solaris), but they cannot compete on demanding generic video (which is your interest). A caveat: there is a significant penalty in the View environment if RDP is chosen rather than PCoIP. The performance with thinlinc to Linux systems is superior to anything under View, provided you can tolerate the cpu and network consumption. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Forcing a DTU to authenticate on a particular server
Adams, Clifford B writes: Is anyone aware of a method whereby I could force a particular DTU to authenticate to a particular server all the time. In my setup, I have two Sun Ray Servers and 24 clients, I want to split the clients up evenly between the two servers and force 12 to always authenticate with one server and the other twelve to authenticate with the other server. Using the Sun Ray load balancing is not an option since I have 4 critical workstations (DTUs) that can't all be down at the same time, which is a possibility when using SRSS load balancing. man utamghadm -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Forcing a DTU to authenticate on a particular server
Adams, Clifford B writes: Is it possible to set the order of preference by which a DTU will connect to a server? For instance, I have the mac address of two DTUs and I want them to connect to Server A as its primary server and the second set of DTUs connect to Server B as their primary servers? Yes, but you will need amgh on both A and B. However (at least in my scenario), machines for B will launch on A if B is down. As Klimov noted, if you have control over DNS or DHCP, there may be more elegant ways to achieve your goal. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] OpenOffice 3.2 font colours and SRSS 4.2
Kevin Doyle writes: I have recently upgraded from SRSS 4.0 to 4.2. I am running Solaris 10 10/09 s10x_u8wos_08a X86 in a FOG Everything is fine except just one thing. When I open a either a new or existing document in Openoffice 3.2 and change the font colour, the colour mapping is wrong [...] Colour mappings seem fine if I open up an image or browse the web. On a similar config, we see blue in certain scenarios involving firefox 4 and flash 10.3. Depending on what other flash objects come into the panel (e.g., during scroll), a flash video will go blue. We don't see this behavior with firefox 3.6.17 and the same flash player. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Multimedia redirection in windows 7 with vmware view
David L. Endicott writes: Kill -USR pid from above You apparently mean USR2. tail /var/dt/Xerrors I'm not seeing the messages in the log. For your example (http://www.neotechsolutions.com/?p=52), I see Display :38.0 Video port Id 65 Compressed: JPEG-D on a similar config (VMWare View 4.5, Rays on Solaris/x86, Win7 32-bit). utcapture reports 0.5-2% packet loss, but this is on a network where we have known problems. prstat reports: PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP 10894 utku7 21M 13M sleep 400 0:03:34 16% uttsc-bin/2 26372 utku7 42M 33M sleep 550 0:02:00 4.1% Xnewt/1 This is on a 2-core VM under ESX on a Sun X4150. We have more packet loss when we run directly on hardware, and we have used the Sun-recommended hires_tick to cut the loss in half. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Multimedia redirection in windows 7 with vmware view
Craig Bender writes: [Rays and multimedia accel in a VMWare View environment] What version of Win7 is this? What VMWare tools are installed on the VM? Any of the multimedia enhancements for View based RDP connections? The only add-on accel software of which I'm aware in this environment are delivered by Sun, and I don't see that these apply to Windows 7: http://wikis.sun.com/display/SRSS4dot2/SRS+5+System+Requirements We run a config similar to Endicott, and the Sun accelerations can be impressive when they work. However, we had sufficient problems (on XP, especially with media player) that we discarded all but the Sun audio driver (to provide audio/video sync). We experience audio/video sync on Windows 7 without the Sun audio driver that is roughly similar to XP with the driver. (We eat more bandwidth with Windows 7, but have not investigated the source.) -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
[SunRay-Users] black-screen on Ray, X11 connection rejected
After applying the Sun Ray software 5.1 bundle on Solaris/x86, Ray 2 are sometimes stalling on a black screen rather than the expected Sun connector for VMWare View. Most of these can be terminated by the Ctrl-Backspace, but sometimes only Ray reset (Ctrl-moon) works. The Ray server console shows: X11 connection rejected because of wrong authentication although I'm uncertain at which point we receive this msg. Admittedly, I've done only superficial examination of logs. I don't know a trigger and so testing has been slow. We did not have this problem with 4.2 Early Access 1. Rays run 4.2_140993-06_2010.10.08.21.53 firmware. I'm uncertain if relevant, but the connections (after the View login and pool selection) are via srwc 2.3 (uttsc). I'm fishing in the hope that other users of the 5.1 bundle have experienced this issue. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SRSS 5.1 and vmware View 4.5
David L. Endicott writes: I added [-N off] to the UTTSC portion of the kiosk arguments (after the --) and it worked! Thanks a million! But now curiosity is killing me. I can't find that parameter in any documentation. What does that parameter do? A good source is http://wikis.sun.com/display/SRWC2dot3/About+Network+Security. The entry from the uttsc man page is: -N on|off Enable or disable enhanced network security (default: on) on - If supported by the remote desktop service, con- nections will use TLS or CredSSP. Otherwise they will use RDP Security. off - RDP Security will be used. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Sunray 3 dropping packets
Ken Mandelberg writes: [packet loss with Ray 3 from Linux Nehalem] What exactly is utcapture displaying as drops, and how can I get more info about where the problem is. We have only Ray 1 and 2, but packet loss of the size you've reported would make our Rays rather unresponsive. In our case, Cisco and Sun blamed eachother, but the basic issue is apparently inability of the switch to buffer properly (as described in the SRSS docs http://wikis.sun.com/display/SRSS4dot2/Home). The full story must be more complicated in our case, as we have a second subnet where things work very well even though the troublesome switches are still involved. For the network with huge loss, we had only a few machines and convinced networking to wire directly to us. This list has a pile of posts on this issue, and no solutions other than to find new switches. You can identify this problem by connecting directly to the server with a cheap switch (we used a $50 Netgear). On Solaris, setting the timer tick was effective in reducing loss. Perhaps not all of this applies to your Ray 3, and we use Solaris (most tests were with the Sun X4200) rather than Linux for the server. We are Gb/s into the switch, but the Ray 1 and 2 are 100 Mb/s. --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] DHCP relay problems on new cisco switch
Dark Bass writes: We have a situation where a customer replaced their old Cisco switches for new Cisco 7010. And now Sunray DTUs on one of the subnets do not get DHCP addresses. While other DHCP clients (PC's) work well on this subnet. 1 other subnet works fine even for Sunrays. Anyone encountered anything like this before ? Yes, although I don't know if the switch was to blame in our case (and I don't know the switch model). University networking sometimes uses dynamic jacks that assign networks by MAC. It appeared to affect all 35 Ray 2, although the Rays would sometimes work. Removing dynamic eliminated the problem. We did not debug sufficiently to understand why Rays are special. As in your report, other systems connected. This list has said that Rays can present multiple MACs at power-on, but networking was skeptical that this was an explanation. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Solaris GDM greeter and SRSS 4.2
CJ Keist writes: So greater question, has anyone on this list gotten both Citrix client and dtlogin to play on the same x86 system together? I haven't tracked recent developments, but we run the ICA Client for Solaris x86 Version 8.50.117422 on a system reporting 10/08 in /etc/release. This system feeds Rays. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Windows Media Player Problem
Matthew Arensberg Wieben writes: There might be, by using group policy preferences, since the audio device selection key resides in HKEY_CURRENT_USER. in reply to Ivar Janmaat: Is there no system wide way of setting Sun Ray RDP Audio for every RDP user? Not elegant, but we do this for XP by group policy, setting HKCU Software\Microsoft\Multimedia\Sound Mapper Value: Playback, REG_SZ, Sun Ray RDP Audio Driver Value: SetupPreferredAudioDevicesCount, REG_DWORD, 0 -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Windows Media Player Problem
Matthew Arensberg Wieben writes: In short, if you have the Sun Ray RDP Audio device selected as the default audio device and play an mp3 (128k,44kHz) through WMP, it locks our sessions up. Has anyone else seen this? We're switching to WinAmp as the default mp3 player for now...it suffers from the 2s audio lag I don't think it is the Sun audio device, but rather the portion that corresponds to your fragment: (you still see the SunMMR 1.0.11 video:none audio:pcm icon in the system tray). I can essentially duplicate your report with 4.2 Early Access 1, using audio-only mp3. However, the lockup isn't always total, and it will eventually cause our sessions via VMware View to disconnect. There are no problems if -M off in uttsc is used. On the bright side, we obtained good audio/video sync using only the Sun audio device (e.g., uttsc has -F off -M off). Admittedly, when the fun accelerations work, they can be significantly better in appearance and load on the server. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Firefox and Flash on Sunrays
Aaron Wilson writes: I used to have the problem where normal flash animations worked. If I went to Youtube or any other site with flash videos, Firefox would become unresponsive for anywhere between 30 seconds to 5 minutes. Generally it lasts for a minute or so. What's the packet loss reported by utcapture? We have packet loss (and apparently worse) on one of our subnets that is not completely explained, and flash animations such as NOAA weather movies can make a Solaris desktop on a Ray nearly useless (at more than 20% loss). However, we do not get the complete stall that you are reporting. If you suspect network issues, a useful test is to run a switch directly off the host and connect the Ray there. We've used a $50 Netgear GS108 for this test. We have a somewhat related issue with video that exploits accelerations described for Rays. With consumer-grade hardware, I can trigger a near-standstill with realplayer when these accelerations are in use. I cannot repeat this on our Sun hardware. Interestingly, packet loss is low (according to utcapture). This issue does not affect flash since flashplayer does not exploit the accelerations. Config: Solaris/x86 on X4200 and X2200, SRSS 4.1 and EA. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SRSS 4.2 Speaker/headphone
Peter Sorensen writes: I have noticed that when I have my headphones plugged into my sunray (sunray 2) the sound comes out of both the headphone and the speaker. Is there any way to fix this or at least disable the speaker? See http://forums.sun.com/thread.jspa?threadID=5407892tstart=15 As a rather ugly workaround, I've considered polling with utset. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SRSS 4.2 Speaker/headphone
Peter Sorensen writes: bash-3.00# /opt/SUNWut/bin/utset -o s=h /etc/opt/SUNWut/basedir/lib/utdevctl: error, no target or session found Run it from the Ray session. To be clear, you will have to continually poll if your users are adjusting volume with the usual Windows control. As soon as users touch the control, both the in-chassis speaker and the headphones go active. We have no issues if the session is to Ray servers running Solaris/x86. rdesktop does not have the problem, but uttsc is required to exploit the enhancements to audio in EA. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SRSS 4.2 Speaker/headphone
Peter Sorensen writes: The only issue is that I use VMware view Kiosk.. We're using the vdm kiosk also. We modified the vdm script under /etc/opt/SUNWkio/sessions/ to set the audio device. However, this won't suffice if users are adjusting volume via the usual Windows control since each adjustment brings the chassis speaker and headphones active (on the Ray 2). Perhaps you can train your users to adjust volume from the Sun-specific keys. The only hack I can see working effectively would be to frequently poll after uttsc is launched, and re-set the device if necessary. This won't be pretty and it apparently will not completely prevent short bursts from the chassis speaker. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] X session terminates when running a specific program
Dr T Crummey writes: This looks like the bug we ran into with Matlab where the Xnewt process (the X server for the SUNRay) core dumps due to a bug in the compiled in MESA library. I think SUN have recognised this as a bug, but as I don't have a support contract for the SUNRays specifically, they won't tell me the ID. On a somewhat related issue, Maple 13 (a symbolic math app) has a new graphics engine that relies on opengl, and it painfully slow on Rays. (It also crashes Secure Global Desktop.) Is there anything to be done that helps? A Maple example illustrating the performance problem is plot3d(x^2+y^2, x=-1..1, y=-1..1) Now try to rotate using the mouse. We're using Maple 12 as a workaround. Config: Solaris/x86 on X4200, displaying Maple from an X4150 running Linux (Maple is not available for Solaris/x86), SRSS 4.1. On the positive side, Solaris/x86 does not appear to have the problem with Matlab experienced by SRSS on Linux. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] usb to parallel converter printing
Craig Bender writes: I've used a couple of different models, one of which was a Belkin device and it worked great. I ran into a couple of issues with certain printers such as an OkiData line printer that our utparlleld just never seemed to work correctly with. But printers like older HP deskjets (non-USB models like a 690C) worked great. It has been some time, but we had troubles with a HP LaserJet 5MP and a USB-to-parallel adapter from APC (not listed in HCL). Printing would work for a short time and then stall. I'm uncertain which part to blame, as we did not debug. If I recall correctly, this was SRSS 3.1 on Solaris/x86. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Flash acceleration works with SRS5
Oliver Braun writes: You might want to check whether the Sun Ray RDP Audio Driver is the default playback device for sound/voice [for Flash acceleration with SRWC early access]. Well...I feel fairly stupid now. Thanks, this indeed fixes audio/video sync. The MMR icon in the tray now indicates both video and audio. -- --Darrel Hankerson hank...@auburn.edu ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Flash acceleration works with SRS5
Lars Tunkrans writes: There is a big difference [with SRSS early access] in the smoothnes of you tube videos run in Internet explorer 8 and the same video run through Firefox 3.5 IE is probably two to three times better in smoothness. But do you get synchronization between audio and video with the new accelerations for Flash? In my tests with Solaris/x86 on an X2200 connecting to Windows XP on a 2.2GHz consumer-grade Pentium 4, Flash is probably improved over the preceding SRWC and the load on systems is less, but there isn't sync (with a Ray 2). If sync is expected, is there a Flash example that is known to work? --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Slow refresh/stripped display updates - an observation
Sean Clarke writes: It has been mentioned a number of times regarding the slow refresh/stripped display updates that occur when using Sun Ray with certain switches at Gbps speed. Our go beyond the Gb issue with Cisco switches. It's true that moving to 100Mb reduces packet loss in some cases, but the multimedia xvideo accelerations can bring our Ray 2 to a near standstill in simple tests with movie clips via realplay. In the realplay case, utcapture shows essentially zero packet loss, unlike the troublesome cases that do not involve xvideo (where we commonly see 8% loss on simple Flash examples when the server is Gb). Like other packet loss problems, we can bypass the problem by inserting a router (a 2-interface ordinary desktop machine) between the Ray and the wall jack. Oddly, the xvideo problem is much less severe on a 2nd Ray server in the same switch stack but running at Gb/s, but we've had unpredictable results with these switches in other tests. What I'd like is a method to gather evidence against the switch in the xvideo example. I do not have admin access to the switch. In the non-xvideo examples, utcapture gives good evidence. Cisco has said there's nothing they can do about the udp buffering problem on these switches. All our problems are solved if we use a $40 Gb Netgear switch as the only switch between the Rays and server. Config: Solaris/x86 on a Sun X4200 and on a consumer-grade core2-quad, SRSS 4.1, Ray 2 (and Ray 1). -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] How to go about selecting a server for a Sun Ray setup
Lars Tunkrans writes: The latest Sun Ray server Software (4.1) have some acceleration to FLASH and MPEG video built in , and the SUN RAY 2 hardware also makes it perform better. The acceleration for Flash in SRSS 4.1 is news to me. I'm aware of the xvideo acceleration (easily seen in tests with realplay), but flash seems unimproved in my simple tests. Further, there was this fragment from Kent Peacock (KP) and William Yang (WY) on 8-June-2009: (WY) ...I noticed that on Craig's blog and I think also somewhere else online (I forget where) it was mentioned that the next release would have improvements for flash video. Are those indeed making it into this release? (KP) Yes, though it will only be from Windows, and running from within IE. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Slow refresh/stripped display updates - an observation
Kent Peacock writes: If you're running on Solaris, did you try adding set hires_tick = 1 to /etc/system and rebooting? Yes, this cuts loss in half in limited non-xvideo tests, but does nothing on the xvideo test with the 100Mb consumer-grade interface. Admittedly, something about our tests don't make sense, as a Sun X4200 on the same switch stack (but at Gb) and to the same Ray 2 works much better. As in many tests, we can bypass the problem by inserting a router between the Ray and wall jack. Is there a way to diagnose the xvideo example that does not involve admin access to the switch? utcapture shows essentially zero packet loss. The Ray 2 is almost stalled, even when the realplay video is stopped. The Ray 2 comes back to life when realplay is terminated. The switch and our side reports 100Mb full duplex. There are no problems if we insert the router mentioned above, even though the traffic to the server is on the same wire. You might tell Cisco that you can replace their switch with a $40 Netgear, and solve the problem. I don't buy that they can't modify their switch firmware to use a better buffering algorithm. We're on the third round with our network people and Cisco, and everyone is aware of the $40 Netgear solution :(. I don't speak to Cisco directly, but I have confidence in our networking personnel. My concern is partly that we are about to deploy 35 classroom Rays, and these will eventually use Flash clips that apparently will be supported by the Sun accelerations in the next SRSS. Thanks. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] How to go about selecting a server for a Sun Ray setup
Craig Bender writes: Just a note of interest, 4.1 on a fast x86 server running a JDS desktop, Flash performance is pretty darn good (baring packet loss issues). Yes, definitely (on our Sun X4200 @2.8GHz running Solaris/x86), although this eats significant resources at the server compared to the redirect-to-Ray acceleration. More importantly for our next deployment: without acceleration, synchronization of audio and video is fairly awful when connecting (with any client, not just Rays) to Windows VMs via vmware view. (SRSS 4.1 has acceleration for some streams via Windows Media player, but we are looking forward to Flash acceleration in the next release. It's hopeless to convince publishers to go cross-platform, so we appear to be dependent on MS-Win.) -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Adobe Reader 9.1 x86
Aaron Wilson writes: In either case [acroread 9 on Solaris/x86 to Ray] takes a good 40-90 seconds before acroread will launch. Usually close to 60 seconds. ... So it leaves me to believe it's something sunray related. I'm leaning towards Xsun, though I wouldn't know how to even confirm or diagnose that. Anyone else noticing anything like this? We see the 1-3 second response reported by Bjoern Rost (on an X4150) if the measurement is for a subsequent launch after acroread is closed, but a cold launch can be 10-30 seconds on an X4200 (two single-core 2.8GHz Opteron) displaying to Xnewt. The issues for us are that cups is required if we want to list printers for users, and crashes on save as text are more common than on Linux/x86. For cups, we point to an existing cups server rather than meddle with the print subsystem on the Ray server. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Off-Topic: Adobe Reader 9.1.0 finally available for Solaris x86!
Eric Bautsch writes: Stupid question: How do you tell CUPS to populate from a central CUPS server? If you mean that you want a simple way to use an existing cups server without changing the stock printing on the Solaris system, then I can describe what I tested. 1. Install SFWcups from a Solaris Companion CD. Versions from 2005 and 2008 worked in my limited testing, but 10_08 does not have cups. 2. cups has changed character set handling (and I don't understand all the issues). We have some versions that don't care, but the target machine does. I changed the file /opt/sfw/cups/lib/locale/en_US/ to have utf-8, but there may be smarter workarounds. Here's the script I tested with acroread (replace ... with your cups server). #!/bin/sh prefix=/opt/sfw/cups LC_ALL=en_US CUPS_SERVER=... PATH=$prefix/bin:$PATH LD_LIBRARY_PATH=$prefix/lib export LC_ALL CUPS_SERVER PATH LD_LIBRARY_PATH exec /opt/Adobe/Reader9/bin/acroread $@ This gives a menu of printers and appears to work, but my testing is limited. As a simpler test, replace the exec line with something like lpstat -p Monitor the cups error log on the cups server when you run the lpstat test. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
[SunRay-Users] utslaunch repeatedly running utdetach
We are experiencing the following on a 4.1 Solaris/x86 Ray server. The keyboard sequence shift-pause to detach succeeds, but runs utdetach repeatedly until the next login, and sufficiently rapid to eat significant resources. We use NSCM, although I'm uncertain if this is relevant. The troublesome machine is updated from SRSS 3.1 (and updated to Solaris 10_08). I'm unable to duplicate on a similar test machine. Running utdetach directly has no problems. I can't find anything on sunsolve, but I'm hoping the issue is known. We do not have ray core services patch 139549 on either system. (Nothing in the patch readme indicates that it would help, however.) -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] utslaunch repeatedly running utdetach
Concerning a flood of utdetach after shift-pause, I wrote: We do not have ray core services patch 139549 on either system. (Nothing in the patch readme indicates that it would help, however.) I suspect I'm mistaken. We have additional problems that are consistent with: 6706040 Xnewt can send auto-repeated keystrokes into a detached session Example: utdetach from the command line can result in repeated Enter into the terminal. I modified utdetach, and shift-pause can cause utslaunch to terminate utdetach and run again (and this repeats rapidly enough to eat significant resources). -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] FYI: Adobe Reader 9.1.0 finally available for Solaris x86!
Stoyan Angelov writes: adobe has finally released a modern port of Adobe Reader for the Solaris x86 platform! The printing support is for cups, and Sun dropped cups from the companion CD for the 10_08 release. Acroread will print without cups, but there's no menu selection of printers. If I recall correctly, the story may be similar for firefox 3 contributed by Sun. Perhaps it's time to switch to cups, but I suspect I'll point acroread to an existing cups server. (In very limited testing, the cups material on our system from 2005 works with 10_08.) What's the smart solution? -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] MMR and VMware View Manager
Brad Lackey writes: Very interesting... I'm not sure that I've even tested this... I would have sworn that it was the TCX extensions built into the VMWare tools, and not the View Agent. I'm not sure if there is any configuration/ installation options of the agent. I confirmed earlier reports by Ross of troubles, but redirection is working with media player 11 on Windows XP Pro service pack 3. I have not done enough tests to be certain that I understand the real problem with earlier tests. I did not return to mediaplayer 10 for verification (Sun says 10 and 11 work). The install of media player 11 was after vmware tools and agent, but this is a system where I've been adding and removing while testing. Tools shows that the Wyse TCX extensions are not available. I believe it is irrelevant, but I did install the 10-March update to SUNWkio-vdm before testing mediaplayer 11. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] slow windows on remote X client
Truls Asheim writes: We are having exactly the same problem which prevents us from switching to Firefox 3. Something which in our opinion is getting quite critical since Mozilla dropped support for Firefox 2 in December. This was in response to: On 09/03/09 09:17, Sammy Atmadja wrote: A lot of newer software (for example firefox 3) relies on having the RENDER extension available on the X-server. When it's not available it tries it tries to compensate for it which results in high network traffic, as it copies a lot of data back and forth. We run a similar config (Solaris/x86 Ray server connecting via ssh to Linux systems for apps that don't run on Solaris). We had a 64-bit Linux system with both 8168 and 8169 realtek interfaces, and the 8168 (presumably the driver) gave significantly better performance. However, we run 8169 in a 32-bit Linux kernel without problems. These are inexpensive Gb interfaces connecting directly to the Ray server (X4200) via Gb switch. Our main concern has been apps such as matlab where life can be painfully slow for remote connections. Fortunately, -Dsun.java2d.pmoffscreen=false gives considerable help in that case. I'd love to hear more general solutions than apply to firefox and acroread. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] MMR and VMware View Manager
On Mar 4, 2009, Nick Ross wrote: All our tests thus far have installed the View agent first, so we'll try installing Sun MMR first to see if that makes a difference. In my tests with the Sun samples, the following fail: 1. vmware tools + agent + Sun MMR 2. vmware tools + agent - agent + Sun MMR + agent 3. Sun MMR + vmware tools + agent There is sound in every case, but no image on a Ray 2 (the Ray 1 works as expected, but this model does not support the associated video acceleration). In terms of TCX, the current version of the View agent doesn't allow us to change or specify components (it's all or nothing), and the TCX extensions (specifically for Wyse) are actually part of the VMware Tools package. By default all our VMs have the TCX extensions excluded. Same here, although I don't know if that means that there are not associcated system changes in our way. I see that Brad Lackey suggests identifying such changes. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SRSS 4.0 and windows 2000 proffesional
Brad Lackey - US-SW Desktop Virtualization Technology Lead brad.lac...@sun.com writes: Why not use VDI 3.0 ? In beta now, set for release at the end of March. We have an RDP server at the virtual hardware level which allows virtually any OS to be presented in this manner. Wile we are only testing a subset of these, Windows 2000 is one of the supported targets. Sun's VirtualBox has a similar-function RDP server; however, it is my understanding that audio (even to Windows XP) will not work this way (that is, to get audio, the customary approach is to bypass the VirtualBox RDP and go directly to XP via RDP). Will audio work with VDI 3.0 in the fashion described? If so, is there hope that general audio/video sync will actually be improved by this facility? (I've tested the Sun multimedia redirection for Rays, but these are targeted too narrowly for the app of interest.) -- --Darrel Hankerson hank...@auburn.edu ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Poor graphics performance redux
Rich Teer writes: If I'm reading that correctly, that's between 20 and 50% packet loss, which would explain a lot! In our experience, Rays are hopeless at this kind of packet loss. I briefly tried swapping out the GBE switch for a Netgear 100 baseT one, but that didn't seem to make much (if any) difference. Any help greatfully received! In our case, we could not resolve problems with Cisco equipment, but instantly solved the problem with inexpensive Netgear Gb/s (GS108) switches. Our host is an X4200 running Solaris 10, and so similar to your config. This switch problem is rather a showstopper for us, since we do not have resources or political influence. Networking here is great, but we are the only campus users of Rays and we have only a handful. For Rays near the server, Networking was willing to insert our switches into their wiring closets. Another workaround we've used: add a router at the target site. We've used ordinary desktop systems for this, with two interfaces (one to the public network, one to the Rays). Not ideal, but it helped in the case of the Ray 1. The timer workaround is reportedly helpful, but we still run SRSS 3.1 and so have not tested. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Graphic performance problems with SunRay 2FS
Craig Bender writes: What does utcapture show? Recently we had a NetGear GigE switch at a trade show. Dropped packets like crazy. Brought in a cheap CompUSA 100 TX switch and all the problems went away. I've seen many posts (and some from me) on packet loss with certain switches. In our case, it was Cisco, and Cisco and Sun blamed eachother. We tried various wiring and port speeds, with some help for the Ray 1 or the Ray 2 (but not both). We routinely had packet loss over 20%, at which point the desktop is mostly unresponsive. For Rays in the same bldg as our server, we inserted $50 Netgear Gb/s switches and worked around the problem. Our server is an X4200. Noted that you did /etc/system work-around, did you reboot? (required) I gather that this is the timer workaround. Does this apply pre-SRSS4 (and on Solaris/x86)? -- --Darrel Hankerson [EMAIL PROTECTED] ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SunRay X server crashes
Philippe Bourdeu d'Aguerre writes: I use Sun Ray Software 4 10/08 on Centos 5.2 X86_64 and I have an application (COMSOL) which repeatedly crashes SunRay X server. There is no message in /var/log/gm/:xx.log Is there anything I can do to avoid this failure or to help debugging the problem ? Probably useless info, but in superficial testing COMSOL on Linux/x86_64 or Solaris/SPARC dsplaying on Rays running from Solaris/x86 works. This is a java thing, but I think comsol uses its own runtime. On the downside, firefox on Solaris/x86 seems to be less stable than on Linux (apparently a flashplayer issue), but we never see crashes of the X server. -- --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] windows connector problem
Andreas Höschler writes: How does windows distinguish devices? All RDP connection are coming from the same machine (Solaris box with SRS and SGD installed). It appears that RDP connections still work from some Sun Rays but not from others. Is another Sun Ray another device for the windows 2003 server? How can it know that? This isn't a direct answer, but Rays are separate devices when using Citrix Metaframe with Windows Terminal Server. We eat CALs per Ray, not per Ray server. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
[SunRay-Users] packet loss, again
We are seeing 30% or more packet loss (from utcapture) on Rays after applying the 27-May-2008 Solaris 10/x86 recommended patch cluster on a Sun X4200. We had recently moved away from Cisco switches since we could not resolve packet loss problems there, and were enjoying exceptional performance with a direct connection to the server via an inexpensive Netgear switch. Users trigger the packet loss problem with flash video or via weather movies. Current packet loss is much worse than anything observed before patching, regardless of wiring. Now of course I regret patching, but the motivation was 127128-11 which addresses a recent security alert. To add to the injury, the sshd patch in the cluster breaks X11 forwarding on machines with only ipv4. We used a workaround of enabling an ipv6 lo0. I don't have evidence that this is related to the packet loss problem, however. I'd be thrilled to hear suggestions or similar experiences. When we were using Cisco switches, we tested the NewtBW throttling mechanism. This helped a little on the Ray 1, but did nothing for our Ray 2. We have SRSS 3.1 with 120880-07. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Firefox on Solaris x86 crashing
William Yang writes: Has anyone else run into issues where Firefox crashes viewing certain webpages? I am running Solaris 10 x86 U4 (8/07). Google Finance (when attempting to view a quote) is the site that most consistently causes segmentation faults in Firefox. The same issue does not occur on a SPARC system (also Solaris 10 U4). If I replace the Sun-shipped Firefox with one of the Solaris builds from the Firefox contributed builds downloads, the crash does not occur. We see this behavior, and we run only the Firefox contributed builds. It's difficult to duplicate, although it often repeats when it occurs. I thought clearing the cache might help, but the evidence isn't convincing. Possibly, we have less trouble if we run from Solaris/SPARC or Linux/x86, but our data here is limited. We run Rays from our X4200, but I don't think this is Ray-specific. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Virtualbox
Stefan Schönherr [EMAIL PROTECTED] writes: I test [virtualbox] last year with Solaris 10 U3 and SRSS 3.1. The RDP-Client (uttsc) can´t connect with the VirtualBox (this time v.1.4) I worked together with the developer from virtualbox to bring it working. With several debugversion and modified VirtualBox-version we can connect a SRSS4 with a VirtualBox 1.5 but the connection wasn't stable. We considered vitualbox for matlab on a Sun X4150 so that we could stay with Solaris, but the one-processor-per-vitualization is a showstopper for us. Now the project is sleeping and i don't know if it is stable. My tests were superficial, but RDP to vitualbox running on Linux (with a 64-bit kernel) was stable, and vitualbox on Solaris/x86 worked very well (but I did not test RDP there). Issues we saw: 1. Installation requires manual labor if running on Linux/x86 with a 64-bit kernel but 32-bit otherwise. 2. The authentication method for MS-Win and RDP connections that uses pam did not work in the scenario of item 1. I suspect this is because we require more 64-bit support (e.g., pam_unix2 for our NIS+ arrangement), but I did not evaluate carefully. I'm aware that Containers for Linux discusses installation of an older matlab, but my tests suggests that making matlab and other apps work smoothly will be difficult. -- --Darrel Hankerson [EMAIL PROTECTED] ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Stream moviefilms with Sun Ray possible?
Scott Serr writes: Youtube sized video is almost acceptable on my little home network. I have a Sun Ray 1 (actually 3) that I was hoping to make into a easy to handle and low power home system for my two kids and wife. It turns out that even small kids like Flash. I'd like to know if anyone has a good experience going to http://www.hotwheels.com ... For me it seems barely usable. This is an example of the type I've reported earlier. The short story seems to be that we have Cisco switches that cannot buffer UDP traffic properly. On your example, packet loss (reported by utcapture) pushes 19%. Much more than this, and the desktop is mostly unresponsive. Try the animations on http://www.ssd.noaa.gov/goes/east/natl.html if you'd like to see what 20+% packet loss does to a Ray 1. If we connect a cheap netgear GS108 switch directly to the X4200, or if we route through a second machine with two cheap interfaces, or if we use Ray 2 (with the X4200 on 100 Mb/s), then packet loss drops below 5%. In addition to packet loss, it seems that the Ray 1 is too slow to process video from www.cnn.com (where packet loss is low, but audio and video is fairly choppy). (It's also necessary to lie to CNN about the user agent to see these videos on Solaris/x86.) Config: Sun X4200 (2 Opteron) running Solaris 10, server and clients in close proximity with very good networking otherwise. We were running the Ray server on SPARC, and I believe the behavior was similar (but we did not test Gb/s on SPARC). -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Gigabit shenanigans
Sean Clarke writes: I have a Sun Ray server based on 2 x dual core Ultra 40. When I attach this to a Sun Ray unit using gigabit ethernet on a Netgear unmanaged gigabit switch I get slow screen redraw performance (with horizontal lines that stalls). Having been monitoring the list I knew an immediate fix would be to force the SR server to 100 Mbit/s and yes, that solved the issue.br If this is packet loss, then our case differs. We have excellent performance when running from an inexpensive Netgear Gb/s switch directly from an X4200, but more than 20% loss with Cisco switches. Going to 100 Mb/s doesn't help the Ray 1, although it makes a difference on Ray 2. We were never able to resolve the problem. Our current workarounds are to use private networks at the Ray site (even though these use the same cabling to reach the Ray server, the packet loss difference is dramatic) or replace with Ray 2. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SRSS 4.0 (09/07) low bandwidth users have poor performance after upgrade
Jimmy Fox writes: The MTU was at 1500. I set it to 1400 and even 1300 but I have the same results. This is a point to point connection without a VPN connection. I would think the 1500 packet would be optimum. This is the size we had before the upgrade. This is perhaps unrelated, but we were never able to solve our latency problems with the Ray 1 in a scenario where there is very good networking and only a single switch stack between the Ray and server. The odd thing is that we get outstanding performance if we use an inexpensive netgear GS108 direct to the server (a Sun X4200), or if we route through a system with Rays running on one of the interfaces and the world out the other. Networking tested 100 Mb/s ports, different switches, etc. Their side sees nothing wrong at the Cisco switches. Our workarounds were to route through other machines (as described above) or move to Ray 2. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] SRSS 4.0 (09/07) low bandwidth users have poor performance after upgrade
Jimmy Fox writes: We don't have bad latency. Just huge amounts of dropped packets. Sorry, I wrote latency when I meant packet loss as measured by utcapture. We can easily trigger 20% packet loss on simple flash animations. At this level, the desktop is largely unresponsive. Darrel makes me think that perhaps because all DTU traffic is funneled through a single interface the server doesn't know how to adjust for the lower network's bandwidth. Since the local LAN is 100Mb, it doesn't know to throttle down for the 3Mb network? I only have one Interface to provide to the DTU traffic so I can't test this theory. Well...we have exceptionally good networking. We tested 100 Mb/s due to recommendations by Sun for somewhat similar problems. In our case, connecting to 100 Mb/s Cisco switch ports helps Ray 2 packet loss, but has little affect on the Ray 1 problem. Our experience where we have very little packet loss on direct connections through inexpensive switches, or routing through another host with Rays on one interface and the world on the other, should explain something. We finally admitted defeat and used workarounds (either a host with two interfaces doing the routing, or replace with Ray 2 and run the X4200 interface to a 100 Mb/s Cisco switch port). -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Sunray Server Choice
Ken Mandelberg writes: So is there a Sunray server that would give faster desktop performance? Would an AMD/Intel server give faster Sunray desktops? Yes, at least compared to a V440 with 1.2GHz processors, the Sun X4200 (Opteron) has better desktop experience. However, you've nicely summarized the difficulties choosing a replacement for our SPARC Ray server. Our compromise was to add a Linux/x86 (and also our Solaris/SPARC) system to run apps that are not available or difficult to build on Solaris/x86. Making apps (e.g., matlab, acrobat reader) launch from the Ray server auto-magically on the other machines is not completely trivial. Matlab and maple are examples that require special hacks to launch across a ssh connection. Also, users expect acroread myfile to work regardless where myfile lives. For temporary files created while browsing, it's easy to tell firefox to place on a shared filesystem, but it's one more issue. A headache we have not solved is when myfile lives on a Ray-connected USB mass storage device. A show-stopper may be sound-related apps such as the flashplayer. Unless there is extra magic, such apps must run on the Ray server. Adobe has released a recent version for Solaris/x86, but Linux versions tend to be first or only. I realize I haven't answered your question. Like you, the reports on this list leave me skeptical about running Rays on Linux. Sun's containers for Linux (brandz) may eventually help for apps such as matlab, but it is not production yet. We really like the Sun X4200 and Solaris/x86, and it has been a very nice upgrade for apps such as the symbolic algebra package magma. However, we have only 10 Rays and we use only the basic non-smartcard login. It may be that our SPARC Ray server had less degradation as load increases, although this is from superficial comparison. -- --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Deploy/Provide new Software to all users?
Theron Dekok writes: When installing applications such as firefox,thunderbird etc you need to create a .desktop entry file in the /usr/share/applications folder. In practice, however, there will be more editing, e.g., /usr/share/{application-registry,mime-info} to make things work smoothly. I have been debating if some explanation of this should be added to the Sun Ray wiki - even though it isn't directly Sun Ray related. One bright spot about CDE was availability of reasonably good printed documentation directed to admins. Sun provides some docs on configuring Gnome (and metacity, etc.), but it is not targeted to admins. I'd love to hear a recommendation. Additions to the wiki would be a welcomed step, however. -- --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] ICA client choices
Brad Lackey writes: [SRS on Linux and use the Citrix client for Linux] is actually looking like a better option considering that Kiosk for Linux will be delivered with the next release. On the other hand, SRS on Linux seems to be the subject of many problem reports on this list, although admittedly some of the problems are due to folks running versions of Linux and Gnome that are not on the official list. I understand the desire to run SRS on Linux, since Solaris/x86 is ignored by vendors. We've tested brandz to run Linux apps, but getting brandz on a SRS looks like an administrative nightmare to me. 3. There is a Citrix Java client. It cannot be installed on the SRS, it has to be loaded dynamically from a web server, presumably one running on the Citrix server. It would probably put high resource demands on the SRS and may be a bit complicated to set up, but it is an option. Don't do it it's very slow and kind of a mess. But it's the only real option for Citrix if SRS is running on Solaris/x86. The Citrix client for Solaris/x86 is ancient. (We also run the SPARC client, but then sound is lost since the Rays run from an X4200.) The native client on SPARC is indeed easier to configure, although it wasn't too painful to configure for the java client. However, our use of Citrix and SRS is basic and involves only a few users. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://node1.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] GigE network to SunRays
Brad Lackey writes: The Sun Ray 2's have a MUCH different (and faster) x86 processor in them. And sufficiently different that our Ray 2 can be made to function where our Ray 1 packet loss is 15% or more (as reported by utcapture). This list has suggested that our problems are possibly duplex mismatch and also to run pingplotter or mtr. Networking says the Cisco switches appear to be fine. For the Rays in this report, there is only a single (Cisco) switch stack between server and client. To use the Ray 2, we moved the server (X4200 running Solaris 10) to a 100 Mb/s line. There are many posts on this very issue, including problems noted by Sun with Gb-to-100Mb downsizing, with the blame sometimes going to switches for dropping UDP packets. The arrangement vastly improves packet loss on the Ray 2, but does not help the Ray 1. To be fair, the Ray 1 functions perfectly if we wire directly to the server via a $50 Netgear GS108 switch. We also have very good performance if we wire through a router (an ordinary Pentium III with two 3com 905 interfaces), even though this routes over the same wire where we have packet loss when connected directly. --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://node1.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Massive network traffic
Partington, David R Mr (NGIT) USAICFH writes: I'm curious what simple video is. They are almost as many types of video as there are ice cream flavors. The original complaint was for weather movies on http://www.ssd.noaa.gov/goes/east/natl.html under Image Loop - Visible. However, this makes the desktop on a Ray 1 mostly unresponsive, so I've also been using http://www.quinlanroad.com/explorethemusic/anancientmuse.asp under E-Card Three. This is a low-bandwidth Macromedia flash video, but it is sufficient to trigger at least 15% packet loss on the Ray 1. Again, if we connect on a private network (a 400 MHz Celeron with two 3com 905 interfaces) that routes over the same wire, or if we use a Ray 2, or if we connect directly to the Sun Ray server, then we have good performance with peak packet loss of a few percent at most. I would make a couple of recommendations for the best performance on Video. First use a dedicated interconnect just for the Sun Rays Unfortunately, we don't control the wire. Networking went as far as to install an older 100 Mb/s switch under the gamble that it would function like our inexpensive Netgear, but I think we'd scrap Rays before pushing this further. and make sure you are using a high quality switch. In fact, it seems what we want is low quality or, rather, an inexpensive switch. We have no problems with $50 Netgear switches (Gb/s or 100 Mb/s); it's the good stuff (e.g., Cisco 3750) that break. I see that you are using Sun Ray 1's. Can you switch solely to Sun Ray 2's since they have a faster processor etc? Our current strategy is to move some Ray 1 to private networks (where we don't have packet loss problems, even though the traffic is eventually on the public wire), or replace with Ray 2. Admittedly, Rays are inexpensive, but it's hard to sell Sun to the dept with this kind of trouble. Thanks for the suggestions. -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://node1.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] UTDEVROOT link incomplete or incorrect?
Hawes writes: I could use some advice on figuring out why my system doesn't have the UTDEVROOT dir/link. $ echo $UTDEVROOT /tmp/SUNWut/sessions/8/unit $ ls -lLd $UTDEVROOT /tmp/SUNWut/sessions/8/unit: No such file or directory SYSTEM INFO Solaris 10 (last patched 8/16/06) SRSS3.1 (3.1_32,REV=2005.08.24.08.55) We had the problem on Solaris 8. The links from /tmp/SUNWut/sessions/ to /tmp/SUNWut/units/ would be missing, which breaks Citrix connections. We did not resolve the issue, instead using a cron job to establish the missing links by running through /tmp/SUNWut/config/dispinfo/. -- --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] GUI for usb removable media drives?
Walter Moore writes: There are a number of things you can do - nothing off the shelf that I know of. Maybe in SRSS 4, eh Sun guys? :) I wrote a script (and I'm happy to share it) that polls utdiskadm every few seconds adds and removes desktop icons for the gnome users... Even simpler would be icons or menu items that open their /tmp/SUNWut/mnt/$USER directory -- at least that way they don't need a command-line (or a good memory for obscure paths). There was a contribution (see below) from Peter Downs that may provide a starting point. --darrel === From: Darrel Hankerson Subject: Re: [SunRay-Users] Automatically show content of usbstick To: sunray-users@filibeto.org Date: Thu, 13 Apr 2006 18:52:47 -0500 (CDT) Concerning Peter Downs wrapper for utodmount (to better handle USB devices), the code (and the suggested revisions I sent) should skip the new portions for the root user. This is an issue when a user leaves devices connected on lock or logout. The root user becomes the owner. Thanks again to Downs. -- --Darrel Hankerson == To: SunRay-Users mailing list sunray-users@filibeto.org Subject: Re: [SunRay-Users] Automatically show content of usbstick From: Darrel Hankerson Date: 12 Apr 2006 15:58:25 -0500 Downs, Peter A. writes: I'm using SRSS 3.1 on Solaris 10 sparc ... I wrote a wrapper around /opt/SUNWut/lib/utdomount to create an icon to on a users GNOME desktop after USB insertion and delete the Icon after USB removal. Thanks! Here's a few comments and suggestions. My tests are on Solaris/x86. 1. The code to translate the device name name=`basename $dev | sed -e 's/s2//'` won't work with my Lexar Jump drive where the device can end disk1p1. (I'm a USB idiot, and I don't know the full story on names.) 2. I've changed USB to be directory of links rather than a link in an attempt to support multiple USB devices. There are corresponding changes in filenames. This makes unmounting less pleasant due to the form of the -p option although it is possible to find the name from the argument. I used a stupid approach that simply checks for existence. I don't know that this is robust. 3. utdomount runs as root. Further, the setuid in chown root:sys /opt/SUNWut/lib/utdomount chmod 4755 /opt/SUNWut/lib/utdomount won't do anything for a shell script utdomount. Running as root will usually not work if home directories are NFS mounted. I've split the script so that most runs as the user. This is a hack, and more scrutiny is certainly required. I'll include it only to show what I tested. Thanks again. ### SUNWut/lib/utdomount #!/bin/sh # Shell wrapper around /opt/SUNWut/lib/utdomount # Adapted from 2005/08/23, Peter Downs, [EMAIL PROTECTED] /opt/SUNWut/lib/utdomount.orig $@ cmdopt=$* id= while [ $# -gt 0 ]; do arg=$1 case $arg in -i) shift id=$1 ;; *) shift ;; esac done if [ $id != ]; then user=`getent passwd $id | cut -d: -f1` if [ $user != ]; then su $user -c /usr/local/sbin/utdomount.sh $cmdopt fi fi ### /usr/local/sbin/utdomount.sh # #!/bin/sh while [ $# -gt 0 ]; do arg=$1 case $arg in -b) shift dev=$1 ;; -i) shift id=$1 ;; -m) shift action=mount ;; -u) shift action=unmount ;; *) shift ;; esac done user=`getent passwd $id | cut -d: -f1` HOME=`getent passwd $id | cut -d: -f6` name=`basename $dev | sed -e 's/s2//' -e 's/p1$//'` icon=$HOME/Desktop/usb_$name.desktop case $action in mount) # Create icon mpath=/tmp/SUNWut/mnt/$user/$name cat $icon EOM [Desktop Entry] Encoding=UTF-8 Name=Removable USB Device $name Exec=nautilus $mpath TryExec=nautilus Icon=gnome-dev-removable Terminal=false Type=Application X-GNOME-Bugzilla-Bugzilla=GNOME X-GNOME-Bugzilla-Product=nautilus X-GNOME-Bugzilla-Component=general EOM if [ ! -d $HOME/USB ]; then mkdir $HOME/USB fi ln -s $mpath $HOME/USB/ ;; unmount) for n in 1 2 3 4 ; do if [ ! -h /tmp/SUNWut/mnt/$user/disk$n ]; then rm $HOME/Desktop/usb_disk$n.desktop rm $HOME/USB/disk$n fi done ;; esac -- Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Linux SRSS impressions
LeBar, Russell J [EMAIL PROTECTED] writes: Up until recently we were heading more towards software thin clients but recent internal events have shifted our focus back to a core contingency of always having some hardware thin clients. This was recently covered on this list. One side argued that Rays are no less labor intensive than other mostly thin clients. Our very limited experience certainly favors Rays. Admittedly, this was with older (1996) HP terminals, but I see many of the same issues on slightly newer NCDs. Using local clients to off-load some of the work never worked smoothly. The X servers were buggy and infrequently updated. Display postscript was an issue. With the Ray, all of the load is indeed on the server, but we never have the headaches seen with the HPs. We are currently using sun ray mobile (NSCM) sessions and use KDE for the desktop (one reason is we have to support multiple simultaneous sessions for users). I'm puzzled why this is desirable, but my limited experience is that Gnome doesn't like this. I recall that the login manager can drop out of mobile mode, and then a login is to a separate session. I don't think this provides a workaround for your goal, however. Given the above, what are your thoughts on using SRSS 3.1 with the SRS4 patch on a supported flavor of Linux? As you've noted, the reports here are not exactly encouraging SRSS+Linux. Further, Solaris has outstanding compilers at the right price. On the downside, important applications (e.g., acroread, matlab, maple) are not available for Solaris/x86, and compile-from-source can be more difficult because building for Solaris/x86 receives less testing. Our solution has been to rely on Linux/x86 and Solaris/SPARC application servers, and automagically feed users. This is not quite as easy as it sounds, since it has to work from within web browsers, isatty type problems, etc. what do people think of using the Solaris x86 version? We moved form Solaris/SPARC to Solaris/x86 with essentially no SRSS issues (even though we moved from Solaris 8 and CDE to Solaris 10 and Gnome). However, we have only a handful of users. Any thoughts on BrandZ to provide a user-space Linux environment? I wish Sun would release a production BrandZ. We tested BrandZ and it was indeed impressive in very limited use (but there were certainly issues). However, using brandz means Solaris Express. For us, this is essentially a show-stopper due to maintenance. -- --Darrel Hankerson [EMAIL PROTECTED] ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Picking the right server
Leigh Porter writes: Maybe I said this earlier but I thought that Solaris 10 x86 could run Linux binaries, in which case, you ought to be able to use the Linux Adobe reader... This is Solaris Containers for Linux. There are snapshots of the development project. The results were promising in our tests earlier this year, but I don't think we could consider this on a production Sun Ray server. We use SPARC and Linux/x86 machines to provide access. This is not without pain, and this issue is a significant obstacle for choosing Solaris/x86. On the positive side, the transition to srss 3 on Solaris/x86 (from srss 2 on SPARC) for the actual Ray software was painless, and life is much faster on the X4200 compared to our Blade 2000. -- --Darrel Hankerson [EMAIL PROTECTED] ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Automatically show content of usbstick
Concerning Peter Downs wrapper for utodmount (to better handle USB devices), the code (and the suggested revisions I sent) should skip the new portions for the root user. This is an issue when a user leaves devices connected on lock or logout. The root user becomes the owner. Thanks again to Downs. -- --Darrel Hankerson ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Automatically show content of usbstick
Downs, Peter A. [EMAIL PROTECTED] writes: I'm using SRSS 3.1 on Solaris 10 sparc ... I wrote a wrapper around /opt/SUNWut/lib/utdomount to create an icon to on a users GNOME desktop after USB insertion and delete the Icon after USB removal. Thanks! Here's a few comments and suggestions. My tests are on Solaris/x86. 1. The code to translate the device name name=`basename $dev | sed -e 's/s2//'` won't work with my Lexar Jump drive where the device can end disk1p1. (I'm a USB idiot, and I don't know the full story on names.) 2. I've changed USB to be directory of links rather than a link in an attempt to support multiple USB devices. There are corresponding changes in filenames. This makes unmounting less pleasant due to the form of the -p option although it is possible to find the name from the argument. I used a stupid approach that simply checks for existence. I don't know that this is robust. 3. utdomount runs as root. Further, the setuid in chown root:sys /opt/SUNWut/lib/utdomount chmod 4755 /opt/SUNWut/lib/utdomount won't do anything for a shell script utdomount. Running as root will usually not work if home directories are NFS mounted. I've split the script so that most runs as the user. This is a hack, and more scrutiny is certainly required. I'll include it only to show what I tested. Thanks again. ### SUNWut/lib/utdomount #!/bin/sh # Shell wrapper around /opt/SUNWut/lib/utdomount # Adapted from 2005/08/23, Peter Downs, [EMAIL PROTECTED] /opt/SUNWut/lib/utdomount.orig $@ cmdopt=$* id= while [ $# -gt 0 ]; do arg=$1 case $arg in -i) shift id=$1 ;; *) shift ;; esac done if [ $id != ]; then user=`getent passwd $id | cut -d: -f1` if [ $user != ]; then su $user -c /usr/local/sbin/utdomount.sh $cmdopt fi fi ### /usr/local/sbin/utdomount.sh # #!/bin/sh while [ $# -gt 0 ]; do arg=$1 case $arg in -b) shift dev=$1 ;; -i) shift id=$1 ;; -m) shift action=mount ;; -u) shift action=unmount ;; *) shift ;; esac done user=`getent passwd $id | cut -d: -f1` HOME=`getent passwd $id | cut -d: -f6` name=`basename $dev | sed -e 's/s2//' -e 's/p1$//'` icon=$HOME/Desktop/usb_$name.desktop case $action in mount) # Create icon mpath=/tmp/SUNWut/mnt/$user/$name cat $icon EOM [Desktop Entry] Encoding=UTF-8 Name=Removable USB Device $name Exec=nautilus $mpath TryExec=nautilus Icon=gnome-dev-removable Terminal=false Type=Application X-GNOME-Bugzilla-Bugzilla=GNOME X-GNOME-Bugzilla-Product=nautilus X-GNOME-Bugzilla-Component=general EOM if [ ! -d $HOME/USB ]; then mkdir $HOME/USB fi ln -s $mpath $HOME/USB/ ;; unmount) for n in 1 2 3 4 ; do if [ ! -h /tmp/SUNWut/mnt/$user/disk$n ]; then rm $HOME/Desktop/usb_disk$n.desktop rm $HOME/USB/disk$n fi done ;; esac ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users
Re: [SunRay-Users] Sun-Rays.org gone???
Brian Knobbs [EMAIL PROTECTED] writes: I'm new, but I'd be willing to help with content. I've been heavily documenting my attempts to get Solaris 10 x86 working, and running SRSS 3.1 on Linux (ubuntu). Possible a wiki? We've just moved from SPARC (srss v2) to Solaris/x86 (srss v3), so I'd love to see x86 notes (on Solaris and Linux). So far, we have not had many surprises, but we have only a very small and simple deployment. -- --Darrel Hankerson [EMAIL PROTECTED] ___ SunRay-Users mailing list SunRay-Users@filibeto.org http://www.filibeto.org/mailman/listinfo/sunray-users