Bug #704
closedit hang both in client and osd
0%
Description
2.in osd0,
- ls /date/osd0 ;hang there,never return.
Files
Updated by longguang yue over 13 years ago
#ls /data/osd0
cat /proc/31361/statck
__mutex_fastpath_lock_retval+0x18/0x1a
vfs_readdir+0x59/0xb2
sys_getdents+0x81/0xd1
system_call_fastpath+0x16/0x1b
0xfffffffffffff
Updated by longguang yue over 13 years ago
at mds0
#cat /proc/469/statck
futex_wait)queue)me+0xc5/0xe4
futex_wait+0x143/0x2f9
do_futex+0x9c/0x852
sys_futex+0x134/0x144
system_cal_fsatpath+0x16/0x1b
0xfffffffffffffff
Updated by longguang yue over 13 years ago
at the first time ,i mkcephfs -a --mkbtrfs.client can mount.ceph,
but if you ls /data/osd0 at osd0-node,it will hang,
mkcephfs can make it work the first time,if it erroe,you have to mkcephfs again.
-----------------------------------------
2011-01-17 13:13:42.103691 7ffe9ccc6740 thread 140731529058048 start
2011-01-17 13:13:42.158882 7ffe9ccc6740 thread 140731520665344 start
2011-01-17 13:13:42.158936 7ffe9ccc6740 thread 140731512272640 start
2011-01-17 13:13:42.159025 7ffe9ccc6740 thread 140731503879936 start
2011-01-17 13:13:42.160186 7ffe9ccc5700 send_observe_requests
2011-01-17 13:13:42.160220 7ffe9ccc5700 mon <- observe pgmap
2011-01-17 13:13:42.160232 7ffe9ccc5700 mon <- observe mdsmap
2011-01-17 13:13:42.160239 7ffe9ccc5700 mon <- observe osdmap
2011-01-17 13:13:42.160245 7ffe9ccc5700 mon <- observe logm
2011-01-17 13:13:42.160253 7ffe9ccc5700 mon <- observe class
2011-01-17 13:13:42.160260 7ffe9ccc5700 mon <- observe monmap
2011-01-17 13:13:42.160266 7ffe9ccc5700 mon <- observe auth
2011-01-17 13:13:42.160273 7ffe9ccc5700 refresh after 150 with same mon
2011-01-17 13:13:42.160334 7ffe9b4c2700 thread 140731502827264 start
2011-01-17 13:13:42.160976 7ffe9b3c1700 reader got ack seq 1
2011-01-17 13:13:42.163976 7ffe9b3c1700 reader got ack seq 9
2011-01-17 13:13:42.164084 7ffe9ccc6740 send_observe_requests
2011-01-17 13:13:42.164098 7ffe9ccc6740 mon <- observe pgmap
2011-01-17 13:13:42.164165 7ffe9ccc6740 mon <- observe mdsmap
2011-01-17 13:13:42.164179 7ffe9ccc6740 mon <- observe osdmap
2011-01-17 13:13:42.164187 7ffe9ccc6740 mon <- observe logm
2011-01-17 13:13:42.164194 7ffe9ccc6740 mon <- observe class
2011-01-17 13:13:42.164202 7ffe9ccc6740 mon <- observe monmap
2011-01-17 13:13:42.164209 7ffe9ccc6740 mon <- observe auth
2011-01-17 13:13:42.164216 7ffe9ccc6740 refresh after 150 with same mon
2011-01-17 13:13:42.167500 7ffe9ccc5700 mon0 > pgmap v100 (latest)
2011-01-17 13:13:42.167497 pg v100: 528 pgs: 528 active+clean; 28707 KB data, 154 MB used, 40803 MB / 40958 MB avail
2011-01-17 13:13:42.168589 7ffe9b3c1700 reader got ack seq 16
2011-01-17 13:13:42.168628 7ffe9ccc5700 mon0 -> mdsmap v6 (latest)
2011-01-17 13:13:42.168628 mds e6: 1/1/1 up {0=up:replay}
2011-01-17 13:13:42.168700 7ffe9ccc5700 mon0 -> osdmap v7 (latest)
2011-01-17 13:13:42.168700 osd e7: 2 osds: 2 up, 2 in
2011-01-17 13:13:42.168798 7ffe9ccc5700 mon0 -> logm v181 (latest)
2011-01-17 13:13:42.168798 log 2011-01-17 13:11:34.262474 mon0 10.28.5.10:6789/0 2 : [INF] mds? 10.28.5.13:6800/26828 up:boot
2011-01-17 13:13:42.168913 7ffe9ccc5700 mon0 -> class v1 (latest)
2011-01-17 13:13:42.168966 7ffe9ccc5700 mon0 -> monmap v1 (latest)
2011-01-17 13:13:42.168965 mon e1: 1 mons at {0=10.28.5.10:6789/0}
2011-01-17 13:13:42.169021 7ffe9ccc5700 mon0 -> auth v5 (latest)
2011-01-17 13:13:42.169107 7ffe9ccc6740 thread 140731503879936 stop
2011-01-17 13:13:42.169165 7ffe9ccc6740 thread 140731502827264 stop
2011-01-17 13:13:42.169208 7ffe9ccc6740 thread 140731529058048 stop
2011-01-17 13:13:42.169245 7ffe9ccc6740 thread 140731512272640 stop
2011-01-17 13:13:42.169269 7ffe9ccc6740 thread 140731520665344 stop
----------------------------------
Updated by longguang yue over 13 years ago
- File dmesg.ceph dmesg.ceph added
this time ,i mkcephfs again.
at first mount.ceph is ok,i cp fine /mnt/ceph ,this is ok,but sync hang,you can ls now.while ceph -w give little msg.
i go to osd1, ls /data/osd1,it hang,but ls osd0 is fine.
[in client] ls /mnt/ceph, is ok, but when cp Newfile to /mnt/ceph , ls /mnt/ceph , it hang!!!!!!
i am try find the point,but have no idea. log is at highest level,but i can not find usable info.
attached file is ,ls's stack,ceph s, dmesg at client,
--ceph -w ----------------
mon0 -> pgmap registered
mon0 -> mdsmap registered
mon0 -> osdmap registered
logm,class,auth registered...........
reader got ack seq 346....
Updated by longguang yue over 13 years ago
it is strange, previous osd0 hang,this time osd1 hang.....
Updated by Sage Weil about 13 years ago
- Status changed from New to Closed
the ceph -s output shows the mds in state 'up:replay'.. that's why the client couldn't mount.