Galera crash on gcache.page. file creation when disk space filled

[Miguel Angel Nieto]

This bug is same one reported here:

https://bugs.launchpad.net/percona-xtradb-cluster/+bug/1488535

and still based on:

https://github.com/codership/galera/issues/324

It is marked as fixed, but this is the log from a user with 5.6.29 (after hundred of lines of new created pages):

2016-08-25 02:15:42 19036 [Note] WSREP: Created page /mysql/mydata/db_data/gcache.page.000360 of size 134217728 bytes
2016-08-25 02:33:42 19036 [Note] WSREP: Created page /mysql/mydata/db_data/gcache.page.000361 of size 134217728 bytes
2016-08-25 02:58:08 19036 [Note] WSREP: Created page /mysql/mydata/db_data/gcache.page.000362 of size 134217728 bytes
2016-08-25 03:21:46 19036 [Note] WSREP: Created page /mysql/mydata/db_data/gcache.page.000363 of size 134217728 bytes
2016-08-25 03:45:57 19036 [Note] WSREP: Created page /mysql/mydata/db_data/gcache.page.000364 of size 134217728 bytes
01:52:04 UTC - mysqld got signal 7 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
We will try our best to scrape up some info that will hopefully help
diagnose the problem, but since we have already crashed,
something is definitely wrong and this may fail.
Please help us make Percona XtraDB Cluster better by reporting any
bugs at https://bugs.launchpad.net/percona-xtradb-cluster

key_buffer_size=134217728
read_buffer_size=2097152
max_used_connections=68
max_threads=502
thread_count=9
connection_count=4
It is possible that mysqld could use up to
key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 2194485 K bytes of memory
Hope that's ok; if not, decrease some variables in the equation.

Thread pointer: 0x0
Attempting backtrace. You can use the following information to find out
where mysqld died. If you see no messages after this, something went
terribly wrong...
stack_bottom = 0 thread_stack 0x40000
/usr/sbin/mysqld(my_print_stacktrace+0x35)[0x907585]
/usr/sbin/mysqld(handle_fatal_signal+0x4b4)[0x66b804]
/lib64/libpthread.so.0[0x3626a0f7e0]
/lib64/libc.so.6(memcpy+0x10e)[0x36266897ae]
/usr/lib64/libgalera_smm.so(_Z22gcs_defrag_handle_fragP10gcs_defragPK12gcs_act_fragP7gcs_actb+0x123)[0x7f1193d2d273]
/usr/lib64/libgalera_smm.so(_Z13gcs_core_recvP8gcs_coreP12gcs_act_rcvdx+0x4bb)[0x7f1193d336eb]
/usr/lib64/libgalera_smm.so(+0x17dfb0)[0x7f1193d3afb0]
/lib64/libpthread.so.0[0x3626a07aa1]
/lib64/libc.so.6(clone+0x6d)[0x36266e893d]
You may download the Percona XtraDB Cluster operations manual by visiting
http://www.percona.com/software/percona-xtradb-cluster/. You may find information
in the manual which will help you identify the cause of the crash.
160825 03:52:05 mysqld_safe Number of processes running now: 0

[Przemek (pmalkowski)]

I haven't tested so many pages created (over 300 reported) due to limited time and disk space, but I wasn't able to repeat the hard crash on PXC 5.6.32, instead, I have this abort only message:

(...)
2016-10-10 16:25:49 21192 [Note] WSREP: Created page /var/lib/mysql/gcache.page.000018 of size 134217728 bytes
2016-10-10 16:26:12 21192 [Note] WSREP: Created page /var/lib/mysql/gcache.page.000019 of size 134217728 bytes
2016-10-10 16:26:34 21192 [Note] WSREP: Created page /var/lib/mysql/gcache.page.000020 of size 134217728 bytes
2016-10-10 16:26:56 21192 [Note] WSREP: Created page /var/lib/mysql/gcache.page.000021 of size 134217728 bytes
2016-10-10 16:27:20 21192 [Note] WSREP: Created page /var/lib/mysql/gcache.page.000022 of size 134217728 bytes
2016-10-10 16:28:25 21192 [ERROR] WSREP: Cannot create new cache page: Requested size 134217728 for '/var/lib/mysql/gcache.page.000023' exceeds available storage space 121303040: 28 (No space left on device)
         at galerautils/src/gu_fdesc.cpp:FileDescriptor():101
2016-10-10 16:28:25 21192 [ERROR] WSREP: gcs/src/gcs_defrag.cpp:gcs_defrag_handle_frag():113: Could not allocate memory for new action of size: 52829111
2016-10-10 16:28:25 21192 [ERROR] WSREP: gcs/src/gcs_core.cpp:core_handle_act_msg():642: Failed to handle action fragment: -12 (Cannot allocate memory)
2016-10-10 16:28:25 21192 [Note] WSREP: gcomm: terminating thread
2016-10-10 16:28:25 21192 [Note] WSREP: gcomm: joining thread
2016-10-10 16:28:25 21192 [Note] WSREP: gcomm: closing backend
2016-10-10 16:28:25 21192 [Note] WSREP: view(view_id(NON_PRIM,2538ad47,1049) memb {
        2538ad47,1
} joined {
} left {
} partitioned {
        86349c69,0
        c39e5108,1
})
2016-10-10 16:28:25 21192 [Note] WSREP: view((empty))
2016-10-10 16:28:25 21192 [Warning] WSREP: 0x1cfaf98 down context(s) not set
2016-10-10 16:28:25 21192 [Warning] WSREP: Send action {(nil), 52829111, TORDERED} returned -107 (Transport endpoint is not connected)
2016-10-10 16:28:25 21192 [Note] WSREP: gcomm: closed
2016-10-10 16:28:25 21192 [Note] WSREP: /usr/sbin/mysqld: Terminated.
161010 16:28:26 mysqld_safe Number of processes running now: 0
161010 16:28:26 mysqld_safe WSREP: not restarting wsrep node automatically
161010 16:28:26 mysqld_safe mysqld from pid file /var/lib/mysql/percona1.pid ended

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PXC-1922