Server File Sharing unstable on mac and windows network on Mavericks and OSX Server 3.0

Florian & rawshooter: great news. So enable guest access along with killing smbd should be the best workaround for now?

Thanks.

Klaus Löfflad

Spezialist für Apple Produkte

http://train4media.de

yes, I'm using it like this since Mavericks was released 😟

But I've escaleted it to Apple support and they are working on this problem.

Regards

Henry

So frustrating, I have about 80 users having a really hard week because of this. We do have a case open with Apple.. Please keep the discussion rolling, specially if you make any adcances!

Possibly to add one thing, if the problems happens on MacMini:

Please check the inside temperatur and CPU-temperature. On my MM it goes from38°C inside to nearly 60°C. The CPU Temp goes upto 90°C

Regards

Henry

Mac Pro 2012

10.9

Server app 3

48 GB RAM

56 TB Fiber array

same issue.

We have 9 Windows XP and one Windows 7 computer that show the same issues.

After a reboot, the Windows machines connect like normal, then randomly, at some point, all the Windows machines will lose the connection (they are all running in a photography lab, with each box running an image capture station, so when one loses connetivity, we know right away)

The dozen or so Macs on the network, connected to the same Mac Pro have no problems. Forcing an SMB connection from the Macs also works fine. Server app lists the PCs as still connected.

Same errors in Console as everyone else....

I'm exaclty in the same situation.

I'm trying to survive until a fix will be released, but it's really hard because users are really frustrating about this issue.

No workarounds, we need a fix very soon.

XServe

10.9

Server 3.0.1

24 GByte RAM

12 TB Raid via Fiber

using

Caching

File sharing

Time Machine

DHCP

DNS

local user accounts (OD didn't work at all after upgrading from 10.6)

running virtual machines with Fusion 6.

I filed a bugreport at http://www.apple.com/feedback/

Klaus Löfflad

Spezialist für Apple Produkte

http://train4media.de

Having the same issue here...

We just replaced a 10+ year old XServe with 10.5 server with a Mac Mini with Mavericks server (hardware died, no choice)

Now, sporadically, smbd goes to 100% usage, everything stops and the mac mini temperature rockets.

Cycling Filesharing or killing smbd temporarily resolves the issue.

Sigh.

(the old xserve chugged a long with basically zero downtime until its PSU finally gave out the other day)

Same Problem here...

after a while, not always same time - smbd hang himself up an go up to 200% and more CPU usage...

MacMini Server, Maverics and Server 3.0

what a pitty

PS: Network is then going down/getting slow - Mac and PC, System doesn't matter.

Definetly there is a probleme with smbd in 10.9 and server 3.0.

But what ist the trigger?

for me it seems that it comes from Windows PC and certain programs.

One program is UPS-Worldship 2013. In general the WordShip program can work with centralized data from OSX server ( 6 month no problem und 10.8). As long as smbd is in normal condition worldship works correctly with 10.9 also.

But it seems that WorldShip can trigger the smbd to "lock".

Any more ideas about possible triggers ??

Regards

Henry

Hey Guys

I noticed something else on SMB connected Macs. The Server shows each Client as inactive, although the user work as usual (I think since connected, the inactive-time runs). Some Clients lost the connection about 3 or 4 Times a day. When they reconnect it works normal.

All AFP Clients are normal.

We have some Windows machines running SQL off the 10.9 server, others only using it as an image repository. All show the same symptoms. The only trigger i can find is (non-guest access) SMB file transfers.

Our system is rote, image files (numbering in the thousands daily) have been the same size for years, database setup has remained unchanged since 10.5 server. There is nothing i can find yet...

Without proper logging from smbd, it is difficult to narrow this down much :/

....just had a passing thought - one of our capture stations is on wifi. we only use it once a week, so it has fallen through the cracks of being checked (aka the server is already rebooted prior to us needing to use it)...

Florian, i noted that too. Server.app doesnt allow you to disconnect SMB clients either, otherwise i was going to see if it helped.

Hi Henry,

for me, a trigger might be, when a user on XP, Win7 oder Win8 starts a program that lies on the server like Office installer, Teamviewer oder banking app. After copying the folder containing the .exe to the desktop I can open it as usual.

Klaus Löfflad

Spezialist für Apple Produkte

http://train4media.de

Andre

Like you we did exhaustive testing, some inpired by our external IT guy, some by Apple.

Largely Apple's advice was useless and in fact switching one of the client PC's to SMB1 only (they must have been reading this blog too) completely took that client off the network and we wasted another half day doing rollbacks before we could get them even registering on the network again. They also got us to do an EDC dump after 'turning logging up' but the turning logging down command that we were given failed to do so and that took our server down completely by filling the whole ssd overnight. We still haven't heard back from Apple re the results from that EDC and it was pretty small as it only took us 90 seconds to replicate the problem.

I've read this blog from start to finish and can say we've tried all of the suggestions, including:

• clean installs of Mavericks and Server 3 (that had no effect);

• setting up new shares on the second ssd on the server (that also had no beneficial effect, but it proved that it was nothing to do with the connected RAID);

• disabling SMB2 & SMB3 (but as stated above that just took that client off the network altogether, apart from retaining internet access);

* setting up a brand new Open Directory (that took everyone off the network - so we had to proceed to the next step);

• setting up new users directly on the server (that seemed to help momentarily, but on reflection it was probably just the fact that we rebooted the server that helped as we were soon back to having the same problems);

We also experinced the problems mentioned above re the ACL's getting corrupted and people loosing read and write permissions to their own files. This didn't just happen with Office files, but our IT guy wrote a few terminal commands relating to the 'temporary items' folder that Office creates in each share point to overwite any client ACL's. That may have helped but I have to admit that very shortly after that we gave up ditched the MacMini Server altogether and went back to using an old xServe running OS10.8.5 after Apple acknowledged that thiswhole debacle is a "known issue" and "their engineers were working on it".

I have to say as a result that I feel seriously aggrieved by this whole experience as we have been working on this issue for over a month with little help from Apple and what supposed 'help' we have received has been useless and merely wasted more of our time and frankly I resent being used as a guinea pig as it has come at an incredible expense to our small office not to mention the dreadful affect it has had on morale.