[Patches] Avoid selftest/autobuild interaction with /tmp

[Stefan Metzmacher]

Hi Amitay,

> Since the TMPDIR change, I have had two random autobuild failures.
> 
> Here's the latest one which fails in talloc:
> 
>   [32/35] abi_check: lib/talloc/bin/default/lib/talloc/libtalloc.so
>   sort: cannot create temporary file in ‘/m/amitay/a/b552249/tmp’: No
> such file or directory
> 
> It appears as if the TMPDIR has been removed prematurely.  Any ideas?

I also got this once, but I don't have any idea yet.

autobuild is running with --verbose now.
If it happens again please collect autobuild.log
and logs.tar.gz and send them to me.

I also got a few flakey ctdb tests lately:

 PASSED  tests/simple/76_ctdb_pdb_recovery.sh
*FAILED* tests/simple/77_ctdb_db_recovery.sh

472/473 tests passed

==========================================================================
--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--
Running test tests/simple/77_ctdb_db_recovery.sh (14:11:41)
--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--
Cluster is HEALTHY
find out which node is recmaster
node 0 is recmaster, disable recmasterrole on node 0
stop node 0
Failed to execute "ctdb stop" on node(s) "0"

*** TEST COMPLETED (RC=1) AT 2016-12-19 14:12:12, CLEANING UP...
Restarting CTDB (scheduled)...
Stopping 3 ctdb daemons...
Starting 3 ctdb daemons...
Waiting for cluster to become ready...
<120|...........|11|
OK
Setting RerecoveryTimeout to 1
Forcing a recovery...
=2|..|
Doing a sync...
ctdb is ready
==========================================================================
TEST FAILED: tests/simple/77_ctdb_db_recovery.sh (status 1) (duration: 62s)
==========================================================================

And another one

 PASSED  tests/simple/31_ctdb_disable.sh
*FAILED* tests/simple/32_ctdb_enable.sh

454/455 tests passed

--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--
Running test tests/simple/32_ctdb_enable.sh (14:05:50)
--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--==--
Cluster is HEALTHY
Selected node 0 with IPs: 192.168.234.1 192.168.234.4.
Disabling node 0
Failed to execute "ctdb disable -n 0" on node(s) "1"

*** TEST COMPLETED (RC=1) AT 2016-12-19 14:06:18, CLEANING UP...
Cluster is UNHEALTHY
DEBUG AT 2016-12-19 14:06:18:
onnode -q 0 ctdb status
Number of nodes:3
pnn:0 127.0.0.11       DISABLED (THIS NODE)
pnn:1 127.0.0.12       OK
pnn:2 127.0.0.13       OK
Generation:2037119717
Size:3
hash:0 lmaster:0
hash:1 lmaster:1
hash:2 lmaster:2
Recovery mode:NORMAL (0)
Recovery master:0
onnode -q 0 onnode all ctdb scriptstatus

>> NODE: /memdisk/kseeger/a/b620888/ctdb/ctdb/tests/var/node.0/ctdbd.socket <<
00.test              OK         0.004 Mon Dec 19 14:06:04 2016

>> NODE: /memdisk/kseeger/a/b620888/ctdb/ctdb/tests/var/node.1/ctdbd.socket <<
00.test              OK         0.001 Mon Dec 19 14:06:10 2016

>> NODE: /memdisk/kseeger/a/b620888/ctdb/ctdb/tests/var/node.2/ctdbd.socket <<
00.test              OK         0.001 Mon Dec 19 14:06:09 2016
Restarting CTDB...
Stopping 3 ctdb daemons...
Starting 3 ctdb daemons...
Waiting for cluster to become ready...
<120|...........|11|
OK
Setting RerecoveryTimeout to 1
Forcing a recovery...
=2|..|
Doing a sync...
ctdb is ready
==========================================================================
TEST FAILED: tests/simple/32_ctdb_enable.sh (status 1) (duration: 55s)
==========================================================================

metze

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: OpenPGP digital signature
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20161219/901d8135/signature.sig>