After I updated the memtable_throughput, I stopped all my writing processes. I did a du /commitlog to find how much was cassandra commitlog at that time. For the three nodes it was around 1.4G each. I waited for about 30 minutes to see whether cassandra flushes things. When I look at du now, it still is around 1.4G. The ls -l on one of the machines shows the following
-rw------- 1 cassandra cassandra 147190162 2011-05-12 17:36 CommitLog-1305221517682.log -rw------- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305221517682.log.header -rw-r--r-- 1 cassandra cassandra 134217815 2011-05-17 00:09 CommitLog-1305590456606.log -rw-r--r-- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305590456606.log.header -rw-r--r-- 1 cassandra cassandra 134217757 2011-05-17 00:18 CommitLog-1305590957399.log -rw-r--r-- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305590957399.log.header -rw-r--r-- 1 cassandra cassandra 134217757 2011-05-17 00:26 CommitLog-1305591492565.log -rw-r--r-- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305591492565.log.header -rw-r--r-- 1 cassandra cassandra 134218024 2011-05-17 00:34 CommitLog-1305591987515.log -rw-r--r-- 1 cassandra cassandra 36 2011-05-17 01:26 CommitLog-1305591987515.log.header -rw-r--r-- 1 cassandra cassandra 137919712 2011-05-17 00:43 CommitLog-1305592441509.log -rw-r--r-- 1 cassandra cassandra 36 2011-05-17 01:26 CommitLog-1305592441509.log.header -rw-r--r-- 1 cassandra cassandra 136446581 2011-05-17 00:59 CommitLog-1305593006344.log -rw-r--r-- 1 cassandra cassandra 36 2011-05-17 01:26 CommitLog-1305593006344.log.header -rw-r--r-- 1 cassandra cassandra 193306617 2011-05-17 01:09 CommitLog-1305594484986.log -rw-r--r-- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305594484986.log.header -rw-r--r-- 1 cassandra cassandra 134986562 2011-05-17 01:21 CommitLog-1305595243108.log -rw-r--r-- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305595243108.log.header -rw-r--r-- 1 cassandra cassandra 134754264 2011-05-17 01:26 CommitLog-1305595537828.log -rw-r--r-- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305595537828.log.header -rw-r--r-- 1 cassandra cassandra 10616832 2011-05-17 01:26 CommitLog-1305595602692.log -rw-r--r-- 1 cassandra cassandra 28 2011-05-17 01:26 CommitLog-1305595602692.log.header There are a couple things that strike me as odd. 1. The first file CommitLog-1305221517682.log is dated 2011/5/12. I wonder why its still lingering around? 2. The times on all the other files range from current to about 1.5 hours ago. Shouldn't this be a smaller list? Thanks! On Mon, May 16, 2011 at 5:44 PM, Sanjeev Kulkarni <sanj...@locomatix.com>wrote: > Hey guys, > I have updated all my column families with 32 as the memtable_throughput. I > will let you know how cassandra behaves. > Thanks! > > > On Mon, May 16, 2011 at 3:52 PM, mcasandra <mohitanch...@gmail.com> wrote: > >> You can try to update column family using cassandra-cli. Try to set >> memtable_throughput to 32 first. >> >> [default@unknown] help update column family; >> update column family Bar; >> update column family Bar with <att1>=<value1>; >> update column family Bar with <att1>=<value1> and <att2>=<value2>...; >> >> Update a column family with the specified values for the given set of >> attributes. Note that you must be using a keyspace. >> >> valid attributes are: >> - column_type: Super or Standard >> - comment: Human-readable column family description. Any string is >> acceptable >> - rows_cached: Number or percentage of rows to cache >> - row_cache_save_period: Period with which to persist the row cache, in >> seconds >> - keys_cached: Number or percentage of keys to cache >> - key_cache_save_period: Period with which to persist the key cache, in >> seconds >> - read_repair_chance: Probability (0.0-1.0) with which to perform read >> repairs on CL.ONE reads >> - gc_grace: Discard tombstones after this many seconds >> - column_metadata: null >> - memtable_operations: Flush memtables after this many operations (in >> millions) >> - memtable_throughput: ... or after this many MB have been written >> - memtable_flush_after: ... or after this many minutes >> - default_validation_class: null >> - min_compaction_threshold: Avoid minor compactions of less than this >> number of sstable files >> - max_compaction_threshold: Compact no more than this number of sstable >> files at once >> - column_metadata: Metadata which describes columns of column family. >> Supported format is [{ k:v, k:v, ... }, { ... }, ...] >> Valid attributes: column_name, validation_class (see comparator), >> index_type (integer), index_name. >> >> >> -- >> View this message in context: >> http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Commitlog-Disk-Full-tp6356797p6370913.html >> Sent from the cassandra-u...@incubator.apache.org mailing list archive at >> Nabble.com. >> > >