Time Machine read extremely slow

[kazooless]

I have several Macs backing up to my TM AFP share. The backups seem to run fine. The AFT share is only shared via AFP with the Time Machine option turned on. When trying to restore from Time Machine it is so slow it times out. When manually mounting the share and trying to mount the disk image TM created, it takes 10's of minutes. When trying to copy the disk image to the local machine it takes days just to get started. I had it tracking at literally just a couple of bits/second.

If I copy a large file to or from the TM share, it behaves properly.

iperf 2 reports roughly 950 Mbits/second in both directions

dd copy reports 1899902788 bytes/sec

So I'm thinking this is just an AFP problem somewhere. How do I troubleshoot this further? Where are the AFP logs? Configs?

Also, it seems to be only with the sparse bundle Time Machine backup files. Regular files can copy to and from the same AFP share with no problem.

Reccomendations?


EDIT:

It's been so long since I've needed to post I forgot to include this information. Is there a CLI command I should run to give more information?

Build FreeNAS-9.10.2-U1 (86c7ef5)
Platform Intel(R) Core(TM) i3-3220T CPU @ 2.80GHz
Memory 8061MB
System Time Fri Mar 03 16:16:59 PST 2017
Uptime 4:16PM up 21 days, 58 mins, 1 user
Load Average 0.13, 0.22, 0.23

It is an Intel Desktop Motherboard so no ECC RAM. It is running off of an MSATA drive on the motherboard. I have 4 WD Red 4TB drives.

 

[joeschmuck]

Reccomendations?

List your system specs per the forum rules and the version of FreeNAS you are running. It could be a hardware limitation or software bug but without this critical data someone would be just shooting in the dark.

 

[kazooless]

List your system specs per the forum rules and the version of FreeNAS you are running. It could be a hardware limitation or software bug but without this critical data someone would be just shooting in the dark.

Sorry about that. Edited original post.

 

[joeschmuck]

How full is your pool? Do you have at least 20% free capacity?

 

[kazooless]

How full is your pool? Do you have at least 20% free capacity?

>datavol 72%
>>datavol 75%
>>>TimeMachine 67%
>>>Home 46%

 

[kazooless]

I should add, clicking on the console at the bottom where I can see only 3 lines, gives me the console window but it never populates. How do I view these logs from the CLI? (I'm ssh'd in)

Here are the last three lines (I am very slowly able to parse the directory structure from Terminal right now with the sparse bundle mounted)

Mar 3 16:06:47 freenas cnid_metad[87852]: main: no volume for path "/mnt/datavol/home"
Mar 3 16:06:47 freenas afpd[87879]: transmit: Request to dbd daemon (volume jeff's home) timed out.
Mar 3 16:06:47 freenas afpd[87879]: afp_openvol(/mnt/datavol/home): Fatal error: Unable to get stamp value from CNID backend

 

[joeschmuck]

I wish I used Time Machine but I don't. I was looking for a hardware issue and right now I suspect it's either a low RAM issue, your pool is very fragmented (yes it's a thing but not something you can fix by normal means), or maybe a software issue.

I guess the last question to ask is: Has this ever worked properly for you? If it has never worked then I'd check to see if there is a bug report for it, at the same time I'd try an earlier version of FreeNAS to see if that makes any difference. If neither work out then I'd file a bug report.

One other thing you can test is to have the computer and FreeNAS system connected directly to each other via Ethernet cable, no router, switch, WiFi, etc... We have found in other situations that throughput issues are related to external hardware and you should check that out as well even though it doesn't look like it would be that. You would be surprised at some of the obscure things we see here.

 

[kazooless]

Holy crap. I spent an hour just trying to traverse the directory structure via terminal of the mounted disk image. All of the sudden it just "let go" and is behaving as one would expect. I really don't know what the heck happened or if it is going to stay this way!

Location of relevant log files would be beneficial if someone could point me in that direction.

 

[kazooless]

After working with this half the night and all morning, I am pretty convinced it is only due to a corrupt disk image. I'm having a really hard time getting it repaired, but I don't think this is a FreeNAS problem now. Thanks for the help.