Yep, same here!
Thank you for the fast fix.
-Jan
On 26.10.2016 23:30, sigsegv wrote:
> The beta version seems to have fixed the problems on my system as well.
>
> Appreciate the fast turnaround time on this!
>
> Justin
> sigsegv
>
> On Wed, Oct 26, 2016 at 1:35 PM, Charles Huber wrote:
>
>> Yup, the beta code seems to fix startup on my ZFS filesystem, thanks!
>>
>> On Wed, Oct 26, 2016 at 3:19 PM, Eric Smith
>> wrote:
>>> We've released a fix that you can try if you're experiencing the problem
>> described in this thread. To test the fix, you will need to opt-in to the
>> steamcmd beta branch. To do this, you have two options:
>>>
>>> 1. run with "-clientbeta publicbeta" on the command line, or
>>>
>>> 2. In the steamcmd install folder there's a "package" directory,
>> make a file named "beta" in there and put this text in it:
>>>
>>> publicbeta
>>>
>>> Just that word on one the top line, nothing else.
>>>
>>> Let me know if you're still having problems after testing the fix.
>> Thanks.
>>>
>>> -Eric
>>>
>>>
>>> -Original Message-
>>> From: hlds_linux-boun...@list.valvesoftware.com [mailto:
>> hlds_linux-boun...@list.valvesoftware.com] On Behalf Of sigsegv
>>> Sent: Wednesday, October 26, 2016 11:38 AM
>>> To: Half-Life dedicated Linux server mailing list
>>> Subject: Re: [hlds_linux] [hlds] Mandatory Team Fortress 2 update
>> released
>>>
>>> [CC'd to John Schoenick: please look into this, or forward it on to
>> whoever works on steamcmd.]
>>>
>>> Yes, I too have run into this problem with steamcmd over the last week
>> or so, as my Linux dedicated server installation is on ZFS.
>>>
>>> This appears to be a false positive from steamcmd: it queries the block
>> size of the filesystem and says "holy crap, 128KB? can't deal with that"
>>> even though it isn't actually a problem since ZFS's recordsize isn't
>> really a fixed block size per se.
>>>
>>> I found that reducing the 'recordsize' property of the ZFS volume from
>> 128K to 512 bytes made the errors go away, but then the ZFS performance
>> became hideously slow (not an unexpected result), so I didn't go forward
>> with that workaround.
>>>
>>> (You can run the command 'stat -f .' on a directory to see what the
>> reported block size is; with recordsize=128K it was 128KB, and with
>>> recordsize=512 it was 512B.)
>>>
>>> For the time being, I found a reasonable workaround to be to mount an
>>> ext4 partition on the steamapps/downloading directory. The game
>> directory itself can still be ZFS with normal 128K recordsize; it's just
>> the downloading directory that steamcmd loses its mind over. It'll copy the
>> updated files over to the ZFS game dir just fine.
>>>
>>> Hope that helps, and I hope Valve fixes this brokenness...
>>>
>>> Justin
>>> (sigsegv)
>>>
>>> On Wednesday, October 26, 2016, Charles Huber
>> wrote:
>>>
Yup, I have the servers installed on a ZoL volume too, thanks for the
confirmation!
On Wed, Oct 26, 2016 at 11:05 AM, Jan >>> >
wrote:
> Hey,
>
> are you using ZFS on linux?
> I had the same problem, steamcmd failed to update the server. It
> works only on my ext4 partition for some reason.
> Maybe it is a combination of ZFS on linux and the fix for the dirty
> cow
> bug: https://dirtycow.ninja/
>
>
> On 26.10.2016 17:08, Charles Huber wrote:
>> Hrm, still startup looping:
>>
>> WARNING: No map specified! Server may not heartbeat.
>> Auto detecting CPU
>> Using default binary: ./srcds_linux Server will auto-restart if
>> there is a crash.
>> Updating server using Steam.
>>
>> Redirecting stderr to '/home/gameserver/Steam/logs/stderr.txt'
>> Looks like steam didn't shutdown cleanly, scheduling immediate
>> update
check
>> [ 0%] Checking for available updates...
>> [] Verifying installation...
>> Steam Console Client (c) Valve Corporation
>> -- type 'quit' to exit --
>> Loading Steam API...Created shared memory when not owner
>> SteamController_Shared_mem OK.
>> login anonymous
>>
>> Connecting anonymously to Steam Public...Logged in OK Waiting for
>> license info...OK force_install_dir ./tf2 app_update 232250
>> validate Update state (0x3) reconfiguring, progress: 0.00 (0 / 0)
>> ../tier1/fileio.cpp (3897) : Assertion Failed: CFileReader::Read
>> must be called with a cubData value that is a multiple of the
>> sector size when using unbuffered IO ../tier1/fileio.cpp (3897) :
>> Assertion Failed: CFileReader::Read must be called with a cubData
>> value that is a multiple of the sector size when using unbuffered
>> IO Update state (0x81) committing, progress: 100.00 (180409744 /
180411440)
>> ../tier1/fileio.cpp (3897) : Assertion Failed: CFileReader::Read
>> must be called with a cubData value that is a multiple of the
>> sector size when using