NAS Posted April 15, 2013 Share Posted April 15, 2013 Thanks RobJ and dgaschk. I'll check the BIOS and plan a RAM upgrade for the near future. There isn't a mainboard available with 16 SATA ports, so we can get rid of SAS, or is there? You are probably joking but just in case consider the replacement issue. The rarer and more bespoke MB you buy the harder and more expensive it is to replace on failure. Choose a common MB and rely on SAS cards. Use any spare money to buy a spare SAS card Link to comment
madshi Posted April 15, 2013 Share Posted April 15, 2013 Nah, wasn't meant to be a joke, because it seems that SAS JBOD support in unRAID/Linux isn't as good as support for "native" MB SATA ports. Anyway, I didn't know that MBs with so many ports were so expensive. Probably doesn't make sense that way... Link to comment
WeeboTech Posted April 15, 2013 Share Posted April 15, 2013 rc10 is not absolutely perfect in this respect, but it is far, far, better than rc11/12 Agreed. I tried rc11 and then rc12. I just gave up and went back to rc10. rc10 is just about as solid as can be for me. I might be nuts, but to me something like would seem like show stopper to a full release. I know many want the full version released and there is now a poll for it, but slow write speeds is one thing but non-functional NFS seems pretty major. I wonder if there is any thought as to why it is having such issues. I believe there was a prior post in response to the NFS stale handles issue. If I remember correctly, it had something to do with the shared filesystem. Link to comment
mrow Posted April 15, 2013 Share Posted April 15, 2013 Nah, wasn't meant to be a joke, because it seems that SAS JBOD support in unRAID/Linux isn't as good as support for "native" MB SATA ports. Anyway, I didn't know that MBs with so many ports were so expensive. Probably doesn't make sense that way... There are tons of people on here running SAS cards in JBOD with no issues. I've been using an SASLP-MV8 for years without a single problem. Added an IBM M1015 that I flashed to IT a few months ago and it's also had no issues... Link to comment
madshi Posted April 16, 2013 Share Posted April 16, 2013 There are tons of people on here running SAS cards in JBOD with no issues. I've been using an SASLP-MV8 for years without a single problem. Added an IBM M1015 that I flashed to IT a few months ago and it's also had no issues... Well, I did have a problem which seems to have been caused by SAS not handling a problematic harddisk correctly, see my logs earlier in this thread (and yes, my SASLP-MV8 already had the latest firmware). Basically web access got lost, the problematic harddisk stopped responding altogether, resulting in the repair failing completely. SAS works fine if there are no problems. But I'm wondering how well it works with certain kinds of problems. This problematic harddisk which brought SAS into serious trouble just reported a couple of media errors when connected to a MB SATA port. Other than those media errors it worked just fine on the MB SATA port. This allowed me to repair my array (had to delete 3 files which were damaged by those media errors). With SAS only, the data on the whole disk might have been unrecoverable... FWIW, I'm not 100% sure whether it was the fault of the SAS controller, the firmware, the drivers, or unRAID 5.0-rc12. I never experienced this problem with 4.7. But then maybe I was just lucky there, I don't know... Link to comment
sambo Posted April 17, 2013 Share Posted April 17, 2013 Hello, webinterface crash here, and cant restart it from cli. Apr 17 08:01:27 Tower emhttp: unRAID System Management Utility version 5.0-rc12a Apr 17 08:01:27 Tower emhttp: Copyright © 2005-2012, Lime Technology, LLC Apr 17 08:01:27 Tower emhttp: Pro key detected, GUID: XXXX-XXXX-XXXX-XXXXXXXXXXXX Apr 17 08:01:27 Tower emhttp: get_config_idx: fopen /boot/config/flash.cfg: No such file or directory - assigning defaults Apr 17 08:01:28 Tower emhttp: diskFsStatus.1 not found Apr 17 08:01:28 Tower kernel: emhttp[17447]: segfault at 0 ip b74fe760 sp bf89ffd0 error 4 in libc-2.11.1.so[b7485000+15c000] Link to comment
PeterB Posted April 17, 2013 Share Posted April 17, 2013 webinterface crash here, and cant restart it from cli. Correct! emhttp was never intended to be started from the cli - you must reboot. Link to comment
Joe L. Posted April 17, 2013 Share Posted April 17, 2013 The segfault means it crashed because something is invading its memory... Not emhttp itself.. Some other product, a plugin most likely is behaving badly.. I had that with plex when I tried that. This is not "windows", programs do not invade other processes memory. Each process has its own memory space. The only possibility is another program using all available memory, leaving none for emhttp when it goes to try to allocate some. (there is a concept of shared memory, but it is not something you can accidentally invade) It could EASILY be a shared library loaded by a plugin that is not compatible with the unRAID kernel. Disable all plugins (anything installing a shared library file ) and see if the issue goes away. Other possibility is bad memory, or memory not set up properly for voltage, clock speed, or timing in the BIOS. Most BIOS set these for you, most get it right, some do not. It must be set for your specific RAM strips make/model. Link to comment
madshi Posted April 17, 2013 Share Posted April 17, 2013 programs do not invade other processes memory. Each process has its own memory space. The same is true in Windows. Link to comment
Joe L. Posted April 17, 2013 Share Posted April 17, 2013 Looking closer, on stock unRAID rc12a, ldd reports this: ldd /usr/local/sbin/emhttp linux-gate.so.1 => (0xb77b9000) libpthread.so.0 => /lib/libpthread.so.0 (0xb779a000) libcrypt.so.1 => /lib/libcrypt.so.1 (0xb7767000) libcrypto.so.0 => /lib/libcrypto.so.0 (0xb761f000) libc.so.6 => /lib/libc.so.6 (0xb74bc000) /lib/ld-linux.so.2 (0xb77ba000) libdl.so.2 => /lib/libdl.so.2 (0xb74b7000) You crashed in a completely different libc.so (possibly one you installed, overwriting the supplied one?) Apr 17 08:01:28 Tower kernel: emhttp[17447]: segfault at 0 ip b74fe760 sp bf89ffd0 error 4 in libc-2.11.1.so[b7485000+15c000] Edit: Based on the last post, I looked on my server. libc.so.6 is indeed linked to libc-2.11.1.so, so that just leaves a memory issue (or a possibly corrupted config/super.dat file) Joe L. Link to comment
boof Posted April 17, 2013 Share Posted April 17, 2013 Looking closer, on stock unRAID rc12a, ldd reports this: ldd /usr/local/sbin/emhttp linux-gate.so.1 => (0xb77b9000) libpthread.so.0 => /lib/libpthread.so.0 (0xb779a000) libcrypt.so.1 => /lib/libcrypt.so.1 (0xb7767000) libcrypto.so.0 => /lib/libcrypto.so.0 (0xb761f000) libc.so.6 => /lib/libc.so.6 (0xb74bc000) /lib/ld-linux.so.2 (0xb77ba000) libdl.so.2 => /lib/libdl.so.2 (0xb74b7000) You crashed in a completely different libc.so (possibly one you installed, overwriting the supplied one?) Apr 17 08:01:28 Tower kernel: emhttp[17447]: segfault at 0 ip b74fe760 sp bf89ffd0 error 4 in libc-2.11.1.so[b7485000+15c000] Joe L. libc.so.6 => /lib/libc.so.6 (0xb74bc000) is usually just a symlink with the 'real' version underneath. You'd have to look at what the symlink points to in each case. libc.so.6 -> libc-2.11.1.so* On my install, though not rc12 and I have addons installed. So pinch of salt. Link to comment
sambo Posted April 17, 2013 Share Posted April 17, 2013 root@Tower:~# ldd /usr/local/sbin/emhttp linux-gate.so.1 => (0xb7764000) libpthread.so.0 => /lib/libpthread.so.0 (0xb7744000) libcrypt.so.1 => /lib/libcrypt.so.1 (0xb7711000) libcrypto.so.0 => /lib/libcrypto.so.0 (0xb75c9000) libc.so.6 => /lib/libc.so.6 (0xb7466000) /lib/ld-linux.so.2 (0xb7765000) libdl.so.2 => /lib/libdl.so.2 (0xb7461000) root@Tower:~# ls -ls /lib/ | grep libc.so.6 0 lrwxrwxrwx 1 root root 14 2013-03-23 21:51 libc.so.6 -> libc-2.11.1.so* simpleFeatures.core.webGUI-1.0.11-i486-1.txz simpleFeatures.log.viewer-1.0.10-i486-1.txz simpleFeatures.system.info-1.0.10-i486-1.txz encfs-1.7.2-i486-1.tgz screen-4.0.3-i486-1.tgz curl-7.20.1-i486-1.txz htop-0.8.3-i486-1sl.tgz nano-2.2.4-i486-1.txz tcl-8.5.9-i486-1.txz Goin to disable & reboot the box this weekend when i will have more free time Link to comment
dgaschk Posted April 17, 2013 Share Posted April 17, 2013 The segfault means it crashed because something is invading its memory... Not emhttp itself.. Some other product, a plugin most likely is behaving badly.. I had that with plex when I tried that. A segmentation fault is caused by a process attempting to access memory that it is not allowed to access. It cannot be caused by another process. If emhttp is seg-faulting then emhttp or one of the libraries it is calling is the issue. Some other product cannot cause emhttp to segfault. If process A attempts to access process B's memory then process A is terminated with a segfault; not process B (e.g., emhttp). If an add-on replaces a library used by emhttp then the use of the wrong library version can cause the segfault. Link to comment
ratmice Posted April 18, 2013 Share Posted April 18, 2013 Most people use CIFS not NFS so theres less pressure to fix. Well on my Mac, SMB/CIFS just sucks. NFS is much faster and much more stable. One super irritating feature is that limited characters available for file naming lead to nonsense names when browsing via SMB, NFS does not have this limitation. When using SMB the finder oftentimes just locks up when you are accessing a few different disks, sometimes requiring restarting the finder or possibly a complete reboot. Fragmented files are also possible if you are performing a copy operation. I assume this is some limitation of OS X implementation of SMB, but I am forced to deal with it. Maybe some other Mac users will comment. TLDR; NFS is an important feature. Link to comment
mrow Posted April 20, 2013 Share Posted April 20, 2013 Most people use CIFS not NFS so theres less pressure to fix. Well on my Mac, SMB/CIFS just sucks. NFS is much faster and much more stable. One super irritating feature is that limited characters available for file naming lead to nonsense names when browsing via SMB, NFS does not have this limitation. When using SMB the finder oftentimes just locks up when you are accessing a few different disks, sometimes requiring restarting the finder or possibly a complete reboot. Fragmented files are also possible if you are performing a copy operation. I assume this is some limitation of OS X implementation of SMB, but I am forced to deal with it. Maybe some other Mac users will comment. TLDR; NFS is an important feature. In response to your last few sentences. Starting with OS X 10.7 Lion Apple was forced to drop Samba and switch to a version of smb they developed in house because the samba team switched their license to GPLv3 which basically makes it impossible for anyone that makes commercial software to include GPLv3 code in their products, seemingly ignoring the fact that by far the largest contributors to FOSS these days are corporations. Yet another initiative by RMS that clearly will make everyone's lives better and all software free. Probably why there hasn't been a rush to embrace it. In any case Windows and Macs should have no problem talking to each other but Macs and Linux boxes after 10.6 via smb is another story whereas before it would have been perfect since they were both using the exact same code to talk to each other. Link to comment
ratmice Posted April 23, 2013 Share Posted April 23, 2013 Should have stated up front that I am still on Snow Leopard, as Lion and Mountain Lion made one of my 'puters useless. Link to comment
c3 Posted April 29, 2013 Share Posted April 29, 2013 Most people use CIFS not NFS so theres less pressure to fix. Well on my Mac, SMB/CIFS just sucks. NFS is much faster and much more stable. One super irritating feature is that limited characters available for file naming lead to nonsense names when browsing via SMB, NFS does not have this limitation. When using SMB the finder oftentimes just locks up when you are accessing a few different disks, sometimes requiring restarting the finder or possibly a complete reboot. Fragmented files are also possible if you are performing a copy operation. I assume this is some limitation of OS X implementation of SMB, but I am forced to deal with it. Maybe some other Mac users will comment. TLDR; NFS is an important feature. In response to your last few sentences. Starting with OS X 10.7 Lion Apple was forced to drop Samba and switch to a version of smb they developed in house because the samba team switched their license to GPLv3 which basically makes it impossible for anyone that makes commercial software to include GPLv3 code in their products, seemingly ignoring the fact that by far the largest contributors to FOSS these days are corporations. Yet another initiative by RMS that clearly will make everyone's lives better and all software free. Probably why there hasn't been a rush to embrace it. In any case Windows and Macs should have no problem talking to each other but Macs and Linux boxes after 10.6 via smb is another story whereas before it would have been perfect since they were both using the exact same code to talk to each other. All true, but you can install Samba on Mac OS X yourself http://www.networkedmediatank.com/showthread.php?tid=65018 Link to comment
mrow Posted April 30, 2013 Share Posted April 30, 2013 Most people use CIFS not NFS so theres less pressure to fix. Well on my Mac, SMB/CIFS just sucks. NFS is much faster and much more stable. One super irritating feature is that limited characters available for file naming lead to nonsense names when browsing via SMB, NFS does not have this limitation. When using SMB the finder oftentimes just locks up when you are accessing a few different disks, sometimes requiring restarting the finder or possibly a complete reboot. Fragmented files are also possible if you are performing a copy operation. I assume this is some limitation of OS X implementation of SMB, but I am forced to deal with it. Maybe some other Mac users will comment. TLDR; NFS is an important feature. In response to your last few sentences. Starting with OS X 10.7 Lion Apple was forced to drop Samba and switch to a version of smb they developed in house because the samba team switched their license to GPLv3 which basically makes it impossible for anyone that makes commercial software to include GPLv3 code in their products, seemingly ignoring the fact that by far the largest contributors to FOSS these days are corporations. Yet another initiative by RMS that clearly will make everyone's lives better and all software free. Probably why there hasn't been a rush to embrace it. In any case Windows and Macs should have no problem talking to each other but Macs and Linux boxes after 10.6 via smb is another story whereas before it would have been perfect since they were both using the exact same code to talk to each other. All true, but you can install Samba on Mac OS X yourself http://www.networkedmediatank.com/showthread.php?tid=65018 Nice, I will have to look in to that. Do you know if this persists after point updates? i.e. When 10.8.4 comes out and I upgrade from 10.8.3, will this persist or will I have to redo it all? Link to comment
zoggy Posted April 30, 2013 Share Posted April 30, 2013 not sure if it matters, but linux 3.9 came out today. is that one of the hold ups for 5.0 ? http://lkml.indiana.edu/hypermail/linux/kernel/1304.3/02009.html http://kernelnewbies.org/LinuxChanges Link to comment
PeterB Posted April 30, 2013 Share Posted April 30, 2013 not sure if it matters, but linux 3.9 came out today. is that one of the hold ups for 5.0 ? No. Link to comment
Techabyte Posted May 5, 2013 Share Posted May 5, 2013 lol cannot download your main site got defaced ...... Link to comment
whiteatom Posted May 5, 2013 Share Posted May 5, 2013 same problem.. just got home and wanted to upgrade.. but no.... Does anyone here have access to fix this? or is this a Tom issue? whiteatom Link to comment
lainie Posted May 5, 2013 Share Posted May 5, 2013 You can still get to the downloads: You can get to the downloads server via this link (Tom has stated in the past the actual download server is a separate server) http://download.lime-technology.com/download/ The most current release is here: http://download.lime-technology.com/download/unRAID%20Server%205.0-rc12a%20AiO.zip Joe L. Link to comment
Techabyte Posted May 6, 2013 Share Posted May 6, 2013 Thanks mate lol seems to be up now anyways Link to comment
Recommended Posts