loss of node will cause shutdown
Hi, i am a newbie in Vertica, and i come across an issue like this -- i supposed installed properly on single node on SuSE 10.4 (x86_64), and when i started database, the db will be shut several minutes later. and i got this message from MC, that is "Loss of node v_test_node0001 will cause shutdown to occur. K=0 total number of nodes=1".
Any advice or suggestions will be much appreciated.
Any advice or suggestions will be much appreciated.
0
Comments
A single node cluster is never k-safe. K-safety means availability of database even in case some of nodes of cluster goes down. But in your case as your database is single node , so it is always at k-safety level K=0, so if a node gets down because of any problem, your database will automatically shutdown.
Vertica recommends K-safety level of 1 (k=1) , which will require at least 3 node cluster. K=1 means vertica is keeping a duplicate copy of vertica projections means data & those copies are used in case a node goes down for recovery of the node OR same copy is accessible by users & which keeps database working.
Hm, that's very interesting... Your process is receiving a SIGFPE; that's an unhandled division-by-zero error.
Is there also a file "ErrorReport.txt" alongside the log? If so, what's in it?
Also, is there anything in particular that you do that tends to cause the database to shut down? (It doesn't look like it from this log snippet...)
Adam
Are you installing on actual hardware or a VM? Are you running any queries before it goes down, or does itgo down whether you have a VSQL session running or not?
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sdc1 488385536 21028164 465823788 5% /
tmpfs 1642984 36 1642948 1% /run
tmpfs 5120 0 5120 0% /run/lock
tmpfs 6623980 0 6623980 0% /run/shm
# df -i
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/sdc1 0 0 0 - /
tmpfs 2053726 24 2053702 1% /run
tmpfs 2053726 1 2053725 1% /run/lock
tmpfs 2053726 2 2053724 1% /run/shm