HAST role failure

[Mike Barnard]

Hi,

finally managed to get the hast resource created with no errors... but I
cannot seem to sync with the secondary. I have this:

hastA# hastctl status all
mail:
  role: primary
  provname: mail

  localpath: /dev/da0s1g
  extentsize: 2097152
  keepdirty: 64
  remoteaddr: tcp4://172.19.254.15
  replication: memsync
  status: degraded
  dirty: 2097152 bytes

tail /var/log/messages:

Dec 14 11:42:32 hastA hastd: [shared] (primary) Unable to receive handshake
header from tcp4://172.19.254.15: Socket is not connected.
Dec 14 11:42:32 hastA hastd: [shared] (primary) Unable to receive handshake
header from tcp4://172.19.254.15: Socket is not connected.
Dec 14 11:42:32 hastA kernel: GEOM_JOURNAL: Journal 1464160566: hast/shared
contains data.
Dec 14 11:42:32 hastA kernel: GEOM_JOURNAL: Journal 1464160566: hast/shared
contains journal.
Dec 14 11:42:32 hastA kernel: GEOM_JOURNAL: Journal hast/shared clean.
Dec 14 11:42:32 hastA kernel: GEOM_JOURNAL: BIO_FLUSH not supported by
hast/shared.
Dec 14 11:42:37 hastA hastd: [shared] (primary) Unable to receive handshake
header from tcp4://172.19.254.15: Socket is not connected.
Dec 14 11:43:12 hastA last message repeated 7 times


On hastB, I get this:

hastB# hastctl status all
mail:
  role: secondary
  provname: mail

  localpath: /dev/da0s1g
  extentsize: 0
  keepdirty: 0
  remoteaddr: tcp4://172.19.254.14

  replication: memsync
  dirty: 0 bytes

tail /var/log/messages:

Dec 14 11:42:32 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:16067: RPC version wrong.
Dec 14 11:42:32 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:64165: RPC version wrong.
Dec 14 11:42:37 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:55836: RPC version wrong.
Dec 14 11:42:42 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:39345: RPC version wrong.
Dec 14 11:42:47 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:14498: RPC version wrong.
Dec 14 11:42:52 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:61407: RPC version wrong.
Dec 14 11:42:57 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:41175: RPC version wrong.
Dec 14 11:43:02 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:42817: RPC version wrong.
Dec 14 11:43:07 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:47832: RPC version wrong.
Dec 14 11:43:12 pollock hastd[3887]: Unable to receive header from tcp4://
172.19.254.14:29129: RPC version wrong.

I cvsuped and rebuilt these two boxes last evening:

uname -a:
FreeBSD hastA 8.1-RELEASE-p2 FreeBSD 8.1-RELEASE-p2 #2: Mon Dec 13 18:33:20
EAT 2010     root at hastA:/
usr/obj/usr/src/sys/KERN.HAST  amd64


I have checked and there was a fix for this, but it seems to be still not
working. Any pointers?

PS: The two servers are identical in hardware and installation.



On Mon, Dec 13, 2010 at 6:38 PM, Mike Barnard <mike.barnardq at gmail.com>wrote:

>
>
> On Mon, Dec 13, 2010 at 5:30 PM, Pawel Jakub Dawidek <pjd at freebsd.org>wrote:
>
>> On Mon, Dec 13, 2010 at 04:33:15PM +0300, Mike Barnard wrote:
>> > Thanks Freddie,
>> >
>> > For some reason, I still get the same error even after destroying the
>> > journal.
>> >
>> > hastctl create shared
>> > [ERROR] [shared] Unable to open /dev/da0s1g: Operation not permitted.
>>
>> Maybe this happens because the operation is not permitted?:)
>>
>
> Isn't da0s1g in use already? For example there is a file system on top
>> of it which is already mounted? If not, please send the output of:
>>
>>
> Yes, da0s1g is, after destroying the journal, mounted and yes has the ufs
> file system on it. and right after sending this, I'll go stick my head in a
> freezer box for good measure of my silliness... sigh
>
>        # sysctl -b kern.geom.confxml
>>
>> I guess there is no need for this since I had grossly overlooked this one
> simple thing #:-s
>
>
> --
> Mike
>
> Of course, you might discount this possibility, but remember that one in
> a million chances happen 99% of the time.
> ------------------------------------------------------------
>



-- 
Mike

Of course, you might discount this possibility, but remember that one in
a million chances happen 99% of the time.
------------------------------------------------------------