limetech Posted June 17, 2013 Author Share Posted June 17, 2013 Bottom line though if you array isn't protected I cannot see the point of requiring initconfig to swap disks about. Yes there are a couple config changes I want to code differently. Most of the restrictions come from the early days when disks were identified by slot id. Link to comment
Donovan Posted June 17, 2013 Share Posted June 17, 2013 can you repair one more thing before this goes to final ? the "TRUST ARRAY" procedure this doesn't work properly for a long time now and it would be nice for it to work properly again... Sorry, what is the "trust array" procedure? I believe he is referring to this old procedure: http://lime-technology.com/wiki/index.php/Make_unRAID_Trust_the_Parity_Drive,_Avoid_Rebuilding_Parity_Unnecessarily This forum post talks about it and what you can do instead with Release 5.0: http://lime-technology.com/forum/index.php?topic=19385.0 Link to comment
sacretagent Posted June 17, 2013 Share Posted June 17, 2013 can you repair one more thing before this goes to final ? the "TRUST ARRAY" procedure this doesn't work properly for a long time now and it would be nice for it to work properly again... Sorry, what is the "trust array" procedure? I believe he is referring to this old procedure: http://lime-technology.com/wiki/index.php/Make_unRAID_Trust_the_Parity_Drive,_Avoid_Rebuilding_Parity_Unnecessarily This forum post talks about it and what you can do instead with Release 5.0: http://lime-technology.com/forum/index.php?topic=19385.0 Yep that's the one but even that version 5 workaround is not working .... at least i never got it to work tried with and without page refresh or are we supposed to do it different way ? as you made a check thingie for trust parity ? Link to comment
nars Posted June 17, 2013 Share Posted June 17, 2013 can you repair one more thing before this goes to final ? the "TRUST ARRAY" procedure this doesn't work properly for a long time now and it would be nice for it to work properly again... Sorry, what is the "trust array" procedure? I believe he is referring to this old procedure: http://lime-technology.com/wiki/index.php/Make_unRAID_Trust_the_Parity_Drive,_Avoid_Rebuilding_Parity_Unnecessarily This forum post talks about it and what you can do instead with Release 5.0: http://lime-technology.com/forum/index.php?topic=19385.0 Yep that's the one but even that version 5 workaround is not working .... at least i never got it to work tried with and without page refresh or are we supposed to do it different way ? as you made a check thingie for trust parity ? Read from Joe L. post at http://lime-technology.com/forum/index.php?topic=19385.msg197921#msg197921 , the only way on latest versions seems using Utils->New Config and then "parity is correct" checkbox, but you need to manually reassign all drives on the right slots... Anyway that is not a thing you want to do often, I think... Link to comment
limetech Posted June 17, 2013 Author Share Posted June 17, 2013 can you repair one more thing before this goes to final ? the "TRUST ARRAY" procedure this doesn't work properly for a long time now and it would be nice for it to work properly again... Sorry, what is the "trust array" procedure? I believe he is referring to this old procedure: http://lime-technology.com/wiki/index.php/Make_unRAID_Trust_the_Parity_Drive,_Avoid_Rebuilding_Parity_Unnecessarily This forum post talks about it and what you can do instead with Release 5.0: http://lime-technology.com/forum/index.php?topic=19385.0 Yep that's the one but even that version 5 workaround is not working .... at least i never got it to work tried with and without page refresh or are we supposed to do it different way ? as you made a check thingie for trust parity ? Do not use that procedure on the wiki - it does not apply to 5.0. As nars pointed out, you use the 'New Config' Utility to erase the configuration. Then go and assign the drives any way you want as long as you assign the correct parity drive. You will notice a check box that says, "Parity is already valid". If you check this it will Start the array and initiate a parity-check instead of a parity-sync. You can cancel the parity-check any time. Link to comment
smakovits Posted June 17, 2013 Share Posted June 17, 2013 RC15 loaded. Removed my intel NIC I was running for some time. will report any issue with realtek nic if they arise. I hope I am past it. Link to comment
tyrindor Posted June 17, 2013 Share Posted June 17, 2013 On both my "problem" servers, RC15 seems to be the first RC that's working with no issues. I believe these mobos/SAS cards were the only major thing still holding us back. 5.0 final here we come... Hope to see 64-bit for 5.1. Link to comment
Lacehim Posted June 17, 2013 Share Posted June 17, 2013 No issues with RC15 for my HP N40L. Link to comment
itimpi Posted June 17, 2013 Share Posted June 17, 2013 Hope to see 64-bit for 5.1. I think it has already been stated that the 5.x series will remain 32-bit with the 6.x series being 64-bit. This means that 6.1 (64-bit) will be functionally identical to 5.1 (32-bit) and will be the first 64-bit release and that both will appear relatively soon after 5.0 final. Link to comment
GFOviedo Posted June 17, 2013 Share Posted June 17, 2013 I just updated from 5B12a to RC15, and so far everything is running well. Link to comment
dikkiedirk Posted June 17, 2013 Share Posted June 17, 2013 Updated to RC15 without copying syslinux.cfg, Mede8er sees NFS shares properly, Parity check in the beginning is very slow between 30-35 MB/sec. Will try later without 4GB memory limit. Unfortunately parity check remains slow, barely 36 MB/s. Where earlier RCs started with close to 100 MB/s and slowly would drop to 60-70 MB/s. Could this be because some disks have only 10 GB space left? Link to comment
garycase Posted June 17, 2013 Share Posted June 17, 2013 Updated to RC15 without copying syslinux.cfg, Mede8er sees NFS shares properly, Parity check in the beginning is very slow between 30-35 MB/sec. Will try later without 4GB memory limit. Unfortunately parity check remains slow, barely 36 MB/s. Where earlier RCs started with close to 100 MB/s and slowly would drop to 60-70 MB/s. Could this be because some disks have only 10 GB space left? No, it's almost certainly because you didn't copy the new syslinux.cfg, which gets rid of that 4GB limit. Stop the parity check; either edit syslinux to kill that 4GB choice or just copy the new one to the server; reboot; and start your parity check up again Link to comment
dikkiedirk Posted June 17, 2013 Share Posted June 17, 2013 Already loaded the new syslinux.cfg. The first 5% of the parity check I get speeds of 30-40 MB/s then it goes up to close to 90 MB/s Link to comment
piotrasd Posted June 17, 2013 Share Posted June 17, 2013 @limetech where we can get source of "kernel 3.9.6-patched (patched with reiserfs sync superblock fix)" or patch for orginal kernel ? Link to comment
darkside40 Posted June 17, 2013 Share Posted June 17, 2013 Updated my Server yesterday. Everything is working fine. Realtek Nic works oob and the Parity check speeds are around 90MB/s just like they should be. Link to comment
Tyler Posted June 17, 2013 Share Posted June 17, 2013 Just had the same network issue I had yesterday happen to me again, here's my previous post. Anyone have any ideas? I'm certainly not a Linux expert, the old "know enough to be dangerous" line comes to mind. But the line "... kernel: NETDEV WATCHDOG: eth0 (r8168): transmit queue 0 timed out" seems to suggest an issue with the Realtek network drivers. So I did some searching and that seems to bolster my suspicion. Here's the syslog from the relevant time for when this dropout happened: Jun 17 21:22:24 Jupiter afpd[12183]: afp_alarm: child timed out, entering disconnected state Jun 17 21:26:07 Jupiter kernel: ------------[ cut here ]------------ Jun 17 21:26:07 Jupiter kernel: WARNING: at net/sched/sch_generic.c:255 dev_watchdog+0x10f/0x18e() Jun 17 21:26:07 Jupiter kernel: Hardware name: To Be Filled By O.E.M. Jun 17 21:26:07 Jupiter kernel: NETDEV WATCHDOG: eth0 (r8168): transmit queue 0 timed out Jun 17 21:26:07 Jupiter kernel: Modules linked in: md_mod w83627ehf hwmon_vid sg i2c_i801 coretemp r8168(O) hwmon i2c_core mpt2sas scsi_transport_sas raid_class ahci libahci [last unloaded: md_mod] Jun 17 21:26:07 Jupiter kernel: Pid: 0, comm: swapper/1 Tainted: G O 3.9.6p-unRAID #2 Jun 17 21:26:07 Jupiter kernel: Call Trace: Jun 17 21:26:07 Jupiter kernel: [] warn_slowpath_common+0x77/0x8e Jun 17 21:26:07 Jupiter kernel: [] ? dev_watchdog+0x10f/0x18e Jun 17 21:26:07 Jupiter kernel: [] ? dev_watchdog+0x10f/0x18e Jun 17 21:26:07 Jupiter kernel: [] warn_slowpath_fmt+0x2e/0x30 Jun 17 21:26:07 Jupiter kernel: [] dev_watchdog+0x10f/0x18e Jun 17 21:26:07 Jupiter kernel: [] ? __netdev_watchdog_up+0x52/0x52 Jun 17 21:26:07 Jupiter kernel: [] call_timer_fn+0x19/0x70 Jun 17 21:26:07 Jupiter kernel: [] run_timer_softirq+0x12b/0x15d Jun 17 21:26:07 Jupiter kernel: [] __do_softirq+0x94/0x151 Jun 17 21:26:07 Jupiter kernel: [] ? sched_clock_idle_wakeup_event+0x11/0x13 Jun 17 21:26:07 Jupiter kernel: [] irq_exit+0x33/0x6c Jun 17 21:26:07 Jupiter kernel: [] smp_apic_timer_interrupt+0x72/0x7f Jun 17 21:26:07 Jupiter kernel: [] apic_timer_interrupt+0x2d/0x34 Jun 17 21:26:07 Jupiter kernel: [] ? cpuidle_wrap_enter+0x2f/0x82 Jun 17 21:26:07 Jupiter kernel: [] cpuidle_enter_tk+0x12/0x14 Jun 17 21:26:07 Jupiter kernel: [] ? disable_cpuidle+0xf/0xf Jun 17 21:26:07 Jupiter kernel: [] cpuidle_enter_state+0xc/0x38 Jun 17 21:26:07 Jupiter kernel: [] cpuidle_idle_call+0x73/0x9b Jun 17 21:26:07 Jupiter kernel: [] cpu_idle+0x46/0x6f Jun 17 21:26:07 Jupiter kernel: [] start_secondary+0x9a/0x9c Jun 17 21:26:07 Jupiter kernel: ---[ end trace 9359dc7f7e64c91a ]--- Jun 17 21:26:07 Jupiter kernel: r8168: eth0: link down Jun 17 21:26:07 Jupiter dhcpcd[1143]: eth0: carrier lost Jun 17 21:26:07 Jupiter avahi-daemon[7578]: Withdrawing address record for 10.10.10.222 on eth0. Jun 17 21:26:07 Jupiter avahi-daemon[7578]: Leaving mDNS multicast group on interface eth0.IPv4 with address 10.10.10.222. Jun 17 21:26:07 Jupiter avahi-daemon[7578]: Interface eth0.IPv4 no longer relevant for mDNS. Jun 17 21:26:11 Jupiter kernel: r8168: eth0: link up Jun 17 21:26:11 Jupiter dhcpcd[1143]: eth0: carrier acquired Jun 17 21:26:11 Jupiter dhcpcd[1143]: eth0: rebinding lease of 10.10.10.222 Jun 17 21:26:11 Jupiter dhcpcd[1143]: eth0: acknowledged 10.10.10.222 from 10.10.10.254 Jun 17 21:26:11 Jupiter dhcpcd[1143]: eth0: checking for 10.10.10.222 Jun 17 21:26:17 Jupiter dhcpcd[1143]: eth0: leased 10.10.10.222 for 604800 seconds Jun 17 21:26:17 Jupiter avahi-daemon[7578]: Joining mDNS multicast group on interface eth0.IPv4 with address 10.10.10.222. Jun 17 21:26:17 Jupiter avahi-daemon[7578]: New relevant interface eth0.IPv4 for mDNS. Jun 17 21:26:17 Jupiter avahi-daemon[7578]: Registering new address record for 10.10.10.222 on eth0.IPv4. Link to comment
jaybee Posted June 17, 2013 Share Posted June 17, 2013 Just tried to update my test server from rc12a, replacing bzimage, bzroot and syslinux.cfg but it gets stuck on reboot stating: chmod: cannot access 'var/log/setup/tmp': no such file or directory. see attachment. I ran a disk checker on the usb stick and that came back ok edit: i have no idea why it's preview is upside down!?, but downloaded it's the right way! Second Edit: Just put 12a back on by overwriting those 3 files and it's working on 12a. Any ideas? I have the same problem and my server takes ages to get to the logon prompt after this. Tom, anyone, what it the recommendation for this problem? Await RC16? Edit something? Modify USB stick files? Link to comment
nars Posted June 17, 2013 Share Posted June 17, 2013 Tom knows it said it will be fixed on next version, anyway this is probably caused by some plugin... in the mean time if you want you either downgrade or you can try to disable your plugins and try to identify what one is causing it. Link to comment
jaybee Posted June 17, 2013 Share Posted June 17, 2013 Tom knows it said it will be fixed on next version, anyway this is probably caused by some plugin... in the mean time if you want you either downgrade or you can try to disable your plugins and try to identify what one is causing it. It's a plugin issue? I only have two things I have installed as I have been waiting for 5 to go final before really jumping on board so I am relatively new to it. I have literally only loaded unmenu and simplefatures that I can think of. I have barely got through the installation/configuration guide so there is nothing "hardcore" running at all. Link to comment
mvdzwaan Posted June 17, 2013 Share Posted June 17, 2013 Tom knows it said it will be fixed on next version, anyway this is probably caused by some plugin... in the mean time if you want you either downgrade or you can try to disable your plugins and try to identify what one is causing it. It's a plugin issue? I only have two things I have installed as I have been waiting for 5 to go final before really jumping on board so I am relatively new to it. I have literally only loaded unmenu and simplefatures that I can think of. I have barely got through the installation/configuration guide so there is nothing "hardcore" running at all. I would classify SF as hardcore.. If it's the amount of people that use it, or the complexity of this plugin, but the number of problem threads I've read on this forum which could be traced back to SF is very large in my opinion. Not to discredit the makers ofcourse Just try it without SF and post your results... Link to comment
nars Posted June 17, 2013 Share Posted June 17, 2013 There was another user with same problem and also with SF... I would try to disable it... or make sure you have the latest version of it... Packages on 'extra' folder doesn't seem the cause of the problem, I did tried it and can't reproduce it, installpkg seems to create /var/log/setup/tmp automatically when trying to install any packages, with no problem at all, I can only guess some plugin attempts to do something on that path... Anyway, Tom said it should be fixed on next version. Link to comment
Point7 Posted June 17, 2013 Share Posted June 17, 2013 No problems with RC15 (running stock config no plugins) Link to comment
stewartwb Posted June 17, 2013 Share Posted June 17, 2013 RC15 working great. Performance improved over previous RCs. Parity check increased to 100MB / sec, back to RC8 levels. I think we're ready for a 5.0 Final - good work, Tom!! Link to comment
needo Posted June 17, 2013 Share Posted June 17, 2013 @limetech where we can get source of "kernel 3.9.6-patched (patched with reiserfs sync superblock fix)" or patch for orginal kernel ? I emailed Jeff (the author of the patch) directly and he sent me to this URL: http://article.gmane.org/gmane.comp.file-systems.reiserfs.general/24261 Thank you. Link to comment
dikkiedirk Posted June 18, 2013 Share Posted June 18, 2013 Unfortunately parity check speed went downhill for me. Where I reached 100+ MB/s at the start in beta12a, I now get about 35 MB/s. I don't know how to improve things, maybe you guys have any tips. I attach a syslog, maybe any of the hardware gurus can tell me what is wrong. syslog-2013-06-18.zip Link to comment
Recommended Posts