July 11, 201213 yr I do backups from unRAID down to a local drive using Chronosync. I often find that my backups have failed because the drive becomes unmounted. In my syslog (on OS X), I see the following: 7/6/12 12:04:16.626 AM KernelEventAgent: tid 00000000 type 'smbfs', mounted on '/Volumes/disk17', from '//redfive@EchoBase-SMB._smb._tcp.local/disk17', dead 7/6/12 12:04:16.626 AM KernelEventAgent: tid 00000000 force unmount //redfive@EchoBase-SMB._smb._tcp.local/disk17 from /Volumes/disk17 7/6/12 12:04:17.000 AM kernel: smb_iod_reconnect: Reconnected share DISK17 with server EchoBase-SMB._smb._tcp.local 7/8/12 3:01:46.000 AM kernel: smb_iod_reconnect: Reconnected share DISK17 with server EchoBase-SMB._smb._tcp.local 7/8/12 3:02:51.000 AM kernel: smb_iod_reconnect: Reconnected share DISK17 with server EchoBase-SMB._smb._tcp.local 7/8/12 3:03:56.000 AM kernel: smb_iod_reconnect: Reconnected share DISK17 with server EchoBase-SMB._smb._tcp.local 7/8/12 3:05:01.000 AM kernel: smb_iod_reconnect: Reconnected share DISK17 with server EchoBase-SMB._smb._tcp.local 7/8/12 3:06:06.000 AM kernel: smb_iod_reconnect: Reconnected share DISK17 with server EchoBase-SMB._smb._tcp.local I use SMB because AFP is still fairly unusable on large arrays even with tweaks. Has anyone else seen anything like this? BTW, on unRAID 5.0 RC4.
July 12, 201213 yr If you google "KernelEventAgent: smbfs dead" there are lots of reports of this. If you don't mind, if you can spend a little gleaning any additional info I'd appreciate it (no time for me to do this at the moment). https://www.google.com/search?q=KernelEventAgent%3A+smbfs+dead&ie=UTF-8 Also, please elaborate on your experience, "AFP is still fairly unusable on large arrays even with tweaks".
July 12, 201213 yr Author If you google "KernelEventAgent: smbfs dead" there are lots of reports of this. If you don't mind, if you can spend a little gleaning any additional info I'd appreciate it (no time for me to do this at the moment). https://www.google.com/search?q=KernelEventAgent%3A+smbfs+dead&ie=UTF-8 Also, please elaborate on your experience, "AFP is still fairly unusable on large arrays even with tweaks". Hi Tom, I did Google these errors as well and found everything from 10.7.4 (which I'm running) changing SMB to reports from years ago. I'm not sure it's unRAID, just curious to see if someone else had seen it. For reference, I mounted the same share via AFP and the backup went through without issue. I should alos mention that when mounting via AFP, I mounted the MP3 User Share. When I mount via SMB, I mount disk17 (where my MP3's reside) so maybe it's an SMB Disk Share issue on Macs? I'll try mounting just the MP3 share via SMB. I did find this Netgear ReadyNAS thread that seems to be what I'm experiencing. It also might be the backup software listing a lot of directories at once then copying them over. As far as I can tell, this *only* disconnects when I'm running Chronosync (which uses rsync as a back end). http://www.readynas.com/forum/viewtopic.php?f=21&t=48120 Regarding AFP: even though I tweak AFP to store the DB files on my cache drive, I find it very slow to connect to AFP shares on my Macs. Sometimes listing directories is an affair that you can start, walk away from, and come back to before it's fully mounted. Sometimes it's fast. Nothing crazy in the syslog (I can post if you want). Also, even though I've told it to connect via 'redfive' (my user account), it will sometimes auto mount as GUEST. I think the slowest part of it is clicking on my unRAID box in the left hand pane of Finder and having it initially list the shares. I've put the share mounts in as a login item and it definitely takes a while that way too. I know that on one of the changelogs you mentioned you were working on some AFP stuff so I figured it would get fixed soon anyway. That's the biggest reason why I've left it alone. Here's the line I use in my GO script to tweak AFP, let me know if it should be fixed. # Set AFP Settings sed -i "c/cnidscheme:dbd/cnidscheme:dbd dbpath:\/mnt/cache/.appledbloc" /etc/netatalk/AppleVolumes.default- Maybe this isn't working because AFP is already started up? Thx again Tom.
Archived
This topic is now archived and is closed to further replies.