[DRBD-user] PingAck did not arrive in time.

Dirk Bonenkamp - ProActive dirk at proactive.nl
Tue Jun 19 08:46:51 CEST 2018


For those who are interested, or those who might stumble on this thread
through Google:

After fiddling around a bit longer without any reliable results, I've
decided to install DRBD8 instead of DRBD9, as Yannis suggested earlier.

Worked straight away, and after a bit of tuning it works awesome.

Kind regards,

D.

On 24-05-18 07:54, Dirk Bonenkamp - ProActive wrote:
> Hello,
>
> Thank you for your suggestion. The MTU is 1500 on both nodes. I had it
> at 9000, but reverted everything to 'normal' to debug this problem.
> Pinging as in your example works fine.
>
> Cheers,
>
> Dirk
>
> On 23-05-18 21:22, Nelson Hicks wrote:
>> Is there any chance this could be an MTU mismatch between the two
>> nodes? If you use ping with varying packet sizes from one node to the
>> other, do they stop working above a specific size? Does ifconfig
>> report the same MTU size for the interface on both nodes?
>>
>> Examples:
>>
>> ifconfig | grep MTU
>>
>> ping -s 500 <other_ip>
>>
>> ping -s 1400 <other_ip>
>>
>> ping -s 1472 <other_ip>
>>
>> ping -s 2000 <other_ip>
>>
>> Thanks,
>>
>> - Nelson Hicks
>>
>>
>>
>>
>> On 05/23/2018 02:07 PM, Dirk Bonenkamp - ProActive wrote:
>>> Hi,
>>>
>>> Thank you for your reply.
>>>
>>> I am / was under the impression that DRBD9 is the new and improved
>>> DRBD, so I figured to use this version. But this is not the case?
>>> Could somebody enlighten me a bit?
>>>
>>> I already have disabled all bonding and other fancy network stuff,
>>> so I'm using  1 nic currently. This doesn't solve anything
>>> unfortunately.
>>>
>>> Kind regards,
>>>
>>> Dirk
>>>
>>> On 23-05-18 14:20, Yannis Milios wrote:
>>>> Two things:
>>>>
>>>> - I would use drbd8 instead of drbd9 for a 2 node setup.
>>>> - I would first test with 1 nic instead of 2.
>>>>
>>>> On Wed, May 23, 2018 at 11:01 AM, Dirk Bonenkamp - ProActive
>>>> <dirk at proactive.nl <mailto:dirk at proactive.nl>> wrote:
>>>>
>>>>     Hi List,
>>>>
>>>>     I'm struggling with a new DRBD9 setup. It's a simple Master/Slave
>>>>     setup.
>>>>     I'm running Ubuntu 16.04 LTS with the DRBD9 packages from the
>>>>     Launchpad PPA.
>>>>
>>>>     I'm running some DRBD8 systems in production for quite some
>>>>     years, so I
>>>>     have some experience. This setup is very similar, the only major
>>>>     difference is that this is DRBD9 and I use LUKS encrypted
>>>>     partitions as
>>>>     backend.
>>>>
>>>>     I keep running into this 'PingAck did not arrive in time.' error,
>>>>     which
>>>>     points to network issues if I am correct (see complete log snippet
>>>>     below). This error occurs when I try to reattach the secondary
>>>> node
>>>>     after a reboot. Initial sync works fine.
>>>>
>>>>     The servers are interconnected with 2 10Gb NICs. I had bonding &
>>>>     jumbo
>>>>     frames configured, but deactivated all this, to no avail. I've
>>>> also
>>>>     stripped the DRBD configuration to the bare minimum (see below).
>>>>
>>>>     I've tested the connection with iperf and some other tools and it
>>>>     seems
>>>>     just fine.
>>>>
>>>>     Could somebody point me in the right direction?
>>>>
>>>>     Thank you in advance, regards,
>>>>
>>>>     Dirk Bonenkamp
>>>>
>>>>     syslog messages:
>>>>
>>>>     May 23 11:31:56 data2 kernel: [  704.111755] drbd: loading
>>>>     out-of-tree
>>>>     module taints kernel.
>>>>     May 23 11:31:56 data2 kernel: [  704.112290] drbd: module
>>>>     verification
>>>>     failed: signature and/or required key missing - tainting kernel
>>>>     May 23 11:31:56 data2 kernel: [  704.127677] drbd: initialized.
>>>>     Version:
>>>>     9.0.14-1 (api:2/proto:86-113)
>>>>     May 23 11:31:56 data2 kernel: [  704.127680] drbd: GIT-hash:
>>>>     62f906cf44ef02a30ce0c148fec223b40c51c533 build by root at data2,
>>>>     2018-05-23
>>>>     09:19:54
>>>>     May 23 11:31:56 data2 kernel: [  704.127683] drbd: registered as
>>>>     block
>>>>     device major 147
>>>>     May 23 11:31:56 data2 kernel: [  704.153565] drbd r0: Starting
>>>> worker
>>>>     thread (from drbdsetup [4495])
>>>>     May 23 11:31:56 data2 kernel: [  704.183031] drbd r0/0 drbd0:
>>>> disk(
>>>>     Diskless -> Attaching )
>>>>     May 23 11:31:56 data2 kernel: [  704.183066] drbd r0/0 drbd0:
>>>> Maximum
>>>>     number of peer devices = 1
>>>>     May 23 11:31:56 data2 kernel: [  704.183293] drbd r0: Method to
>>>>     ensure
>>>>     write ordering: flush
>>>>     May 23 11:31:56 data2 kernel: [  704.183308] drbd r0/0 drbd0:
>>>>     drbd_bm_resize called with capacity == 273437203064
>>>>     May 23 11:31:58 data2 kernel: [  706.508228] drbd r0/0 drbd0:
>>>> resync
>>>>     bitmap: bits=34179650383 words=534057038 pages=1043081
>>>>     May 23 11:31:58 data2 kernel: [  706.508234] drbd r0/0 drbd0:
>>>>     size = 127
>>>>     TB (136718601532 KB)
>>>>     May 23 11:31:58 data2 kernel: [  706.508236] drbd r0/0 drbd0:
>>>>     size = 127
>>>>     TB (136718601532 KB)
>>>>     May 23 11:32:10 data2 kernel: [  717.890420] drbd r0/0 drbd0:
>>>>     recounting
>>>>     of set bits took additional 1256ms
>>>>     May 23 11:32:10 data2 kernel: [  717.890435] drbd r0/0 drbd0:
>>>> disk(
>>>>     Attaching -> Outdated )
>>>>     May 23 11:32:10 data2 kernel: [  717.890439] drbd r0/0 drbd0:
>>>>     attached
>>>>     to current UUID: 244DD61D2781DF44
>>>>     May 23 11:32:10 data2 kernel: [  717.918473] drbd r0 data1:
>>>> Starting
>>>>     sender thread (from drbdsetup [4544])
>>>>     May 23 11:32:10 data2 kernel: [  717.922534] drbd r0 data1: conn(
>>>>     StandAlone -> Unconnected )
>>>>     May 23 11:32:10 data2 kernel: [  717.922820] drbd r0 data1:
>>>> Starting
>>>>     receiver thread (from drbd_w_r0 [4498])
>>>>     May 23 11:32:10 data2 kernel: [  717.922973] drbd r0 data1: conn(
>>>>     Unconnected -> Connecting )
>>>>     May 23 11:32:10 data2 kernel: [  718.421219] drbd r0 data1:
>>>>     Handshake to
>>>>     peer 1 successful: Agreed network protocol version 113
>>>>     May 23 11:32:10 data2 kernel: [  718.421229] drbd r0 data1:
>>>> Feature
>>>>     flags enabled on protocol level: 0xf TRIM THIN_RESYNC WRITE_SAME
>>>>     WRITE_ZEROES.
>>>>     May 23 11:32:10 data2 kernel: [  718.421259] drbd r0 data1:
>>>> Starting
>>>>     ack_recv thread (from drbd_r_r0 [4550])
>>>>     May 23 11:32:10 data2 kernel: [  718.424095] drbd r0: Preparing
>>>>     cluster-wide state change 1205605755 (0->1 499/146)
>>>>     May 23 11:32:10 data2 kernel: [  718.437172] drbd r0: State change
>>>>     1205605755: primary_nodes=2, weak_nodes=FFFFFFFFFFFFFFFC
>>>>     May 23 11:32:10 data2 kernel: [  718.437185] drbd r0: Aborting
>>>>     cluster-wide state change 1205605755 (12ms) rv = -22
>>>>     May 23 11:32:12 data2 kernel: [  719.896223] drbd r0: Preparing
>>>>     cluster-wide state change 445952355 (0->1 499/146)
>>>>     May 23 11:32:12 data2 kernel: [  719.896498] drbd r0: State change
>>>>     445952355: primary_nodes=2, weak_nodes=FFFFFFFFFFFFFFFC
>>>>     May 23 11:32:12 data2 kernel: [  719.896508] drbd r0: Committing
>>>>     cluster-wide state change 445952355 (0ms)
>>>>     May 23 11:32:12 data2 kernel: [  719.896541] drbd r0 data1: conn(
>>>>     Connecting -> Connected ) peer( Unknown -> Primary )
>>>>     May 23 11:32:12 data2 kernel: [  719.912186] drbd r0/0 drbd0
>>>> data1:
>>>>     drbd_sync_handshake:
>>>>     May 23 11:32:12 data2 kernel: [  719.912198] drbd r0/0 drbd0
>>>>     data1: self
>>>>    
>>>> 244DD61D2781DF44:0000000000000000:0000000000000000:0000000000000000
>>>>     bits:52035 flags:20
>>>>     May 23 11:32:12 data2 kernel: [  719.912207] drbd r0/0 drbd0
>>>>     data1: peer
>>>>    
>>>> E38BE51FE782EAE0:244DD61D2781DF44:934CAB8662DF0410:E555BDC58E528356
>>>>     bits:53162 flags:20
>>>>     May 23 11:32:12 data2 kernel: [  719.912214] drbd r0/0 drbd0
>>>> data1:
>>>>     uuid_compare()=-2 by rule 50
>>>>     May 23 11:32:12 data2 kernel: [  719.912248] drbd r0/0 drbd0
>>>> data1:
>>>>     pdsk( DUnknown -> UpToDate ) repl( Off -> WFBitMapT )
>>>>     May 23 11:32:32 data2 kernel: [  740.397026] drbd r0 data1:
>>>>     PingAck did
>>>>     not arrive in time.
>>>>     May 23 11:32:32 data2 kernel: [  740.397121] drbd r0 data1: conn(
>>>>     Connected -> NetworkFailure ) peer( Primary -> Unknown )
>>>>     May 23 11:32:32 data2 kernel: [  740.397131] drbd r0/0 drbd0
>>>> data1:
>>>>     pdsk( UpToDate -> DUnknown ) repl( WFBitMapT -> Off )
>>>>     May 23 11:32:32 data2 kernel: [  740.397176] drbd r0 data1:
>>>>     ack_receiver
>>>>     terminated
>>>>     May 23 11:32:32 data2 kernel: [  740.397182] drbd r0 data1:
>>>>     Terminating
>>>>     ack_recv thread
>>>>     May 23 11:32:32 data2 kernel: [  740.458608] drbd r0 data1:
>>>>     Connection
>>>>     closed
>>>>     May 23 11:32:32 data2 kernel: [  740.458650] drbd r0 data1: conn(
>>>>     NetworkFailure -> Unconnected )
>>>>     May 23 11:32:32 data2 kernel: [  740.458688] drbd r0 data1:
>>>>     Restarting
>>>>     receiver thread
>>>>     May 23 11:32:32 data2 kernel: [  740.458723] drbd r0 data1: conn(
>>>>     Unconnected -> Connecting )
>>>>
>>>>     resources:
>>>>
>>>>     resource r0 {
>>>>             on data1 {
>>>>                     device    /dev/drbd0;
>>>>                     disk      /dev/mapper/mapper_secure;
>>>>                     address 172.16.11.21:7789
>>>> <http://172.16.11.21:7789>;
>>>>                     meta-disk internal;
>>>>             }
>>>>             on data2 {
>>>>                     device    /dev/drbd0;
>>>>                     disk      /dev/mapper/mapper_secure;
>>>>                     address 172.16.11.22:7789
>>>> <http://172.16.11.22:7789>;
>>>>                     meta-disk internal;
>>>>             }
>>>>     }
>>>>
>>>>     drbd configuration:
>>>>
>>>>     global {
>>>>             usage-count yes;
>>>>     }
>>>>
>>>>     common {
>>>>             #handlers {
>>>>             #        fence-peer "/usr/lib/drbd/crm-fence-peer.9.sh
>>>>     <http://crm-fence-peer.9.sh>";
>>>>             #        after-resync-target
>>>>     "/usr/lib/drbd/crm-unfence-peer.9.sh
>>>> <http://crm-unfence-peer.9.sh>";
>>>>             #}
>>>>             #disk {
>>>>             #        on-io-error detach;
>>>>             #       disk-barrier no;
>>>>             #       disk-flushes no;
>>>>             #       al-extents 3833;
>>>>             #        c-plan-ahead 7;
>>>>             #        c-fill-target 2M;
>>>>             #        c-min-rate 80M;
>>>>             #        c-max-rate 720M;
>>>>             #}
>>>>             net {
>>>>                     protocol C;
>>>>                     #fencing resource-only;
>>>>                     #cram-hmac-alg sha1;
>>>>                     #verify-alg sha1;
>>>>                     #shared-secret 1e69dc721fd2e65368ae3ba1e5929979;
>>>>                     #after-sb-0pri disconnect;
>>>>                     #after-sb-1pri disconnect;
>>>>                     #after-sb-2pri disconnect;
>>>>                     #max-buffers    8000;
>>>>                     #max-epoch-size 8000;
>>>>                     #sndbuf-size 0;
>>>>                     #rcvbuf-size 2048k;
>>>>             }
>>>>     }
>>>>
>>>>
>>>>
>>>>     _______________________________________________
>>>>     drbd-user mailing list
>>>>     drbd-user at lists.linbit.com <mailto:drbd-user at lists.linbit.com>
>>>>     http://lists.linbit.com/mailman/listinfo/drbd-user
>>>>     <http://lists.linbit.com/mailman/listinfo/drbd-user>
>>>>
>>>>
>>>
>>> -- 
>>> ProActive Software
>>> Dirk Bonenkamp
>>> CTO         <https://www.proactive-software.com>
>>> Phone: +31 (0)23 54 222 99
>>> Mobile: +31 (0)6 250 787 93     Richard Holkade 9
>>> 2033 PZ Haarlem
>>> LinkedIn <http://linkd.in/1V6egnk> Facebook
>>> <http://bit.ly/FBProActive> YouTube <http://bit.ly/1Mc23L9>
>>> www.proactive.nl <https://www.proactive.nl>
>>>
>>>
>>>
>>> _______________________________________________
>>> drbd-user mailing list
>>> drbd-user at lists.linbit.com
>>> http://lists.linbit.com/mailman/listinfo/drbd-user
>>
>> _______________________________________________
>> drbd-user mailing list
>> drbd-user at lists.linbit.com
>> http://lists.linbit.com/mailman/listinfo/drbd-user
>
> -- 
> ProActive Software
> Dirk Bonenkamp
> CTO 	  	<https://www.proactive-software.com>
> Phone: +31 (0)23 54 222 99
> Mobile: +31 (0)6 250 787 93 	Richard Holkade 9
> 2033 PZ Haarlem
> LinkedIn <http://linkd.in/1V6egnk>  Facebook
> <http://bit.ly/FBProActive>  YouTube <http://bit.ly/1Mc23L9>
> www.proactive.nl <https://www.proactive.nl>
>
>
>
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user

-- 
ProActive Software
Dirk Bonenkamp
CTO 	  	<https://www.proactive-software.com>
Phone: +31 (0)23 54 222 99
Mobile: +31 (0)6 250 787 93 	Richard Holkade 9
2033 PZ Haarlem
LinkedIn <http://linkd.in/1V6egnk>  Facebook
<http://bit.ly/FBProActive>  YouTube <http://bit.ly/1Mc23L9>
www.proactive.nl <https://www.proactive.nl>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20180619/020c382c/attachment-0001.htm>


More information about the drbd-user mailing list