February 28, 201214 yr So I've eventually got everything set up (I've been waiting on various parts, still waiting for my new flash drive so I can buy a license and add more disks though) and I've started the array for the first time and set up some user shares. It's now performing the Parity Sync. Now, I wasn't planning on copying loads of data over straight away, but was curious as to exactly how much a parity sync will affect performance and also wanted to determine that I'd got my directory split level setting correct. Anyway, what seems to happen is that it starts of quite fast for about 5 seconds, then it's suddenly either really slow or freezes, I can't tell. So I try to stop the transfer which both Windows copy and Teracopy seem reluctant to do, and it seems to take ages - I'm talking minutes here, just to stop the copy. I then try to go to the directory and it seems to freeze the samba share, and at some point the WebUI freezes. I think the web UI freezes when I try to go into the subdirectory I was creating/writing to from the WebUI. unMenu does still work however. The final, incredibly odd part to this is that if I then log in via telnet or ssh, navigate to the directory in /mnt/[whatever], and then try to list, the list never completes and just hangs. The only solution really was to powerdown, then start up again and let it perform its parity sync. I'm running 5b14 (as I wasn't sure what the most stable 5b8+ beta was so just went for the latest) and I'm using simple features as the web UI. However, I wouldn't have thought using a different web UI would have this much of an impact. Now I get that transfers should be very slow when synching the parity - that's fine, but I didn't think it would cause a load of things to freeze. Anyone else experience this?
February 28, 201214 yr So I've eventually got everything set up (I've been waiting on various parts, still waiting for my new flash drive so I can buy a license and add more disks though) and I've started the array for the first time and set up some user shares. It's now performing the Parity Sync. Now, I wasn't planning on copying loads of data over straight away, but was curious as to exactly how much a parity sync will affect performance and also wanted to determine that I'd got my directory split level setting correct. Anyway, what seems to happen is that it starts of quite fast for about 5 seconds, then it's suddenly either really slow or freezes, I can't tell. So I try to stop the transfer which both Windows copy and Teracopy seem reluctant to do, and it seems to take ages - I'm talking minutes here, just to stop the copy. I then try to go to the directory and it seems to freeze the samba share, and at some point the WebUI freezes. I think the web UI freezes when I try to go into the subdirectory I was creating/writing to from the WebUI. unMenu does still work however. The final, incredibly odd part to this is that if I then log in via telnet or ssh, navigate to the directory in /mnt/[whatever], and then try to list, the list never completes and just hangs. The only solution really was to powerdown, then start up again and let it perform its parity sync. I'm running 5b14 (as I wasn't sure what the most stable 5b8+ beta was so just went for the latest) and I'm using simple features as the web UI. However, I wouldn't have thought using a different web UI would have this much of an impact. Now I get that transfers should be very slow when synching the parity - that's fine, but I didn't think it would cause a load of things to freeze. Anyone else experience this? Have you looked in the system log for any clues? (hint... we can not look for you, since you did not bother to attach one) There have been tons of people with no issues at all, and others with hardware issues, others with just too much I/O for their system to keep up at any speed. What tests have you performed so far? Is it your network? or your router? or your source PC? or unRAID that is the bottleneck. Attempting to stop a window's copy leaves the Window's PC in a inconsistent state. I've seen where re-booting it cured a lot of ills.
February 29, 201214 yr Author Ah sorry I was dozing off when I posted that and forgot to include them! There is a slight problem with them in that the timestamps are way off, for some reason my system time reset itself at some point. There are a load of disk read errors from 4am today, but that will have all happened after the problem occurred and I rebooted.. Also one of the hard drive didn't have the healthies preclear results (though the SMART test came back as OK), but either way the disk with the failures wasn't the one that I was actually writing to when the problem occurred. I have been getting a load of warning emails saying that one of my disks is disabled or invalid, but I think this is because I (for some stupid reason) set it as the parity and then precleared it - I assume that will go away once the parity sync completes. Anyway, here are the syslog files, where the dated ones are from /boot/logs and the undated one (syslog.txt) is from /var/log. Just to point out that for some reason, the email notification system prints out ALL your details into the syslog, including your email address AND password - I really think that should be sorted. I don't know if it's a "Simple Features" thing or a standard unRAID thing, but I had to scan through and delete all my info from the syslog. In terms of your questions at the end - I don't *think* the network or my PC are the bottlenecks. I just tried copying a load of video files to the flash share and that didn't seem to cause the issue. Also - after seeing the stupid number of errors on the Seagate drive, I've decided to remove it from the array. OK so it just happened again: Feb 29 08:20:08 Tower emhttp: Start array... Feb 29 08:20:08 Tower kernel: mdcmd (13): start NEW_ARRAY Feb 29 08:20:08 Tower kernel: unraid: allocating 18360K for 1280 stripes (3 disks) Feb 29 08:20:08 Tower kernel: md1: running, size: 1953514552 blocks Feb 29 08:20:08 Tower kernel: md2: running, size: 976762552 blocks Feb 29 08:20:08 Tower emhttp: shcmd (647): udevadm settle Feb 29 08:20:08 Tower emhttp: shcmd (648): /usr/local/sbin/emhttp_event array_started Feb 29 08:20:08 Tower emhttp_event: array_started Feb 29 08:20:08 Tower emhttp: Mounting disks... Feb 29 08:20:08 Tower emhttp: shcmd (649): mkdir /mnt/disk2 Feb 29 08:20:08 Tower emhttp: shcmd (650): mkdir /mnt/disk1 Feb 29 08:20:08 Tower emhttp: shcmd (651): set -o pipefail ; mount -t reiserfs -o user_xattr,acl,noatime,nodiratime /dev/md2 /mnt/disk2 |& logger Feb 29 08:20:08 Tower emhttp: shcmd (652): set -o pipefail ; mount -t reiserfs -o user_xattr,acl,noatime,nodiratime /dev/md1 /mnt/disk1 |& logger Feb 29 08:20:08 Tower kernel: mdcmd (14): check CORRECT Feb 29 08:20:08 Tower kernel: md: recovery thread woken up ... Feb 29 08:20:08 Tower kernel: md: recovery thread syncing parity disk ... Feb 29 08:20:08 Tower kernel: REISERFS warning (device md2): sh-2021 reiserfs_fill_super: can not find reiserfs on md2 Feb 29 08:20:08 Tower kernel: REISERFS (device md1): found reiserfs format "3.6" with standard journal Feb 29 08:20:08 Tower kernel: REISERFS (device md1): using ordered data mode Feb 29 08:20:08 Tower kernel: reiserfs: using flush barriers Feb 29 08:20:08 Tower kernel: REISERFS (device md1): journal params: device md1, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30 Feb 29 08:20:08 Tower kernel: REISERFS (device md1): checking transaction log (md1) Feb 29 08:20:08 Tower logger: mount: wrong fs type, bad option, bad superblock on /dev/md2, Feb 29 08:20:08 Tower logger: missing codepage or helper program, or other error Feb 29 08:20:08 Tower logger: In some cases useful info is found in syslog - try Feb 29 08:20:08 Tower logger: dmesg | tail or so Feb 29 08:20:08 Tower logger: Feb 29 08:20:08 Tower emhttp: _shcmd: shcmd (651): exit status: 32 Feb 29 08:20:08 Tower emhttp: disk2 mount error: 32 Feb 29 08:20:08 Tower emhttp: shcmd (653): rmdir /mnt/disk2 Feb 29 08:20:08 Tower kernel: REISERFS (device md1): Using r5 hash to sort names Feb 29 08:20:08 Tower emhttp: shcmd (654): chmod 770 '/mnt/disk1' Feb 29 08:20:08 Tower emhttp: shcmd (655): chown nobody:users '/mnt/disk1' Feb 29 08:20:08 Tower emhttp: shcmd (656): mkdir /mnt/user Feb 29 08:20:08 Tower emhttp: shcmd (657): /usr/local/sbin/shfs /mnt/user -disks 2 -o noatime,big_writes,allow_other,default_permissions,use_ino Feb 29 08:20:08 Tower emhttp: shcmd (658): /usr/local/sbin/emhttp_event disks_mounted Feb 29 08:20:08 Tower emhttp_event: disks_mounted Feb 29 08:20:08 Tower emhttp: shcmd (659): :>/etc/samba/smb-shares.conf Feb 29 08:20:09 Tower kernel: md: using 1536k window, over a total of 1953514552 blocks. Feb 29 08:20:09 Tower emhttp: Restart SMB... Feb 29 08:20:09 Tower emhttp: shcmd (660): killall -HUP smbd Feb 29 08:20:09 Tower emhttp: shcmd (661): ps axc | grep -q rpc.mountd Feb 29 08:20:09 Tower emhttp: _shcmd: shcmd (661): exit status: 1 Feb 29 08:20:09 Tower emhttp: shcmd (662): /usr/local/sbin/emhttp_event svcs_restarted Feb 29 08:20:09 Tower emhttp_event: svcs_restarted Feb 29 08:20:25 Tower emhttp: shcmd (663): set -o pipefail ; mkreiserfs -q /dev/md2 |& logger Feb 29 08:20:25 Tower logger: mkreiserfs 3.6.21 (2009 www.namesys.com) Feb 29 08:20:25 Tower logger: Feb 29 08:21:37 Tower emhttp: shcmd (664): mkdir /mnt/disk2 Feb 29 08:21:37 Tower emhttp: shcmd (665): set -o pipefail ; mount -t reiserfs -o user_xattr,acl,noatime,nodiratime /dev/md2 /mnt/disk2 |& logger Feb 29 08:21:37 Tower kernel: REISERFS (device md2): found reiserfs format "3.6" with standard journal Feb 29 08:21:37 Tower kernel: REISERFS (device md2): using ordered data mode Feb 29 08:21:37 Tower kernel: reiserfs: using flush barriers Feb 29 08:21:37 Tower kernel: REISERFS (device md2): journal params: device md2, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max trans age 30 Feb 29 08:21:37 Tower kernel: REISERFS (device md2): checking transaction log (md2) Feb 29 08:21:38 Tower kernel: REISERFS (device md2): Using r5 hash to sort names Feb 29 08:21:38 Tower kernel: REISERFS (device md2): Created .reiserfs_priv - reserved for xattr storage. Feb 29 08:21:39 Tower emhttp: resized: /mnt/disk2 Feb 29 08:21:39 Tower emhttp: shcmd (666): chmod 770 '/mnt/disk2' Feb 29 08:21:39 Tower emhttp: shcmd (667): chown nobody:users '/mnt/disk2' Feb 29 08:21:39 Tower emhttp: shcmd (668): :>/etc/samba/smb-shares.conf Feb 29 08:21:39 Tower emhttp: Restart SMB... Feb 29 08:21:39 Tower emhttp: shcmd (669): killall -HUP smbd Feb 29 08:21:39 Tower emhttp: shcmd (670): ps axc | grep -q rpc.mountd Feb 29 08:21:39 Tower emhttp: _shcmd: shcmd (670): exit status: 1 Feb 29 08:21:39 Tower emhttp: shcmd (671): /usr/local/sbin/emhttp_event svcs_restarted Feb 29 08:21:39 Tower emhttp_event: svcs_restarted Feb 29 08:25:58 Tower kernel: mdcmd (15): nocheck Feb 29 08:25:58 Tower kernel: md: md_do_sync: got signal, exit... Feb 29 08:25:59 Tower kernel: md: recovery thread sync completion status: -4 syslogs.zip
February 29, 201214 yr Simplefeatures is in debug mode. Turn off debug mode to have it stop adding email and password to the logs. I suggest you run stock unRAID until your comfortable that it is working correctly.
February 29, 201214 yr Disk 2 is failing. It's SMART report indicates over 200 reallocated sectors. It should be replaced. RMA if possible.
February 29, 201214 yr Author Ah thanks, didn't realise it was in debug mode, fixed that at least If you got the drive failing thing from the attached logs, then yeah I was aware that there were a LOT of failures and I've taken it out, pretended it never existed, and put my Samsung 1tb hard drive in. However, it's happened since then and the only log to show while it was happening was the log I put in the actual post. I just looked at the disk attributes for this other disk I'm using, and Raw read error rate's row is: 1 Raw Read Error Rate 0x000f 100 100 051 Pre-fail Always Never 0 I thought that a raw value of 0 was a good result - why does it show Pre-fail?
February 29, 201214 yr Author Hah, oops, I thought it was flagging it as "er, you'd better watch this". I'm kinda new to this SMART stuff.
Archived
This topic is now archived and is closed to further replies.