RE: Booting Cassandra v0.7.0 on Windows: rename failed
Hi, The bug report can be found at: https://issues.apache.org/jira/browse/CASSANDRA-1790 Regards, Ramon From: Jonathan Ellis [mailto:jbel...@gmail.com] Sent: maandag 29 november 2010 16:09 To: user Subject: Re: Booting Cassandra v0.7.0 on Windows: rename failed Please report a bug at https://issues.apache.org/jira/browse/CASSANDRA On Mon, Nov 29, 2010 at 2:49 AM, Ramon Rockx r.ro...@asknow.nl wrote: Hi, Recently I downloaded Cassandra v0.7.0 rc1. When I try to run cassandra it ends with the following logging: INFO 09:17:30,044 Enqueuing flush of memtable-locationi...@839514767(643 bytes, 12 operations) INFO 09:17:30,045 Writing memtable-locationi...@839514767(643 bytes, 12 operations) ERROR 09:17:30,233 Fatal exception in thread Thread[FlushWriter:1,5,main] java.io.IOError: java.io.IOException: rename failed of d:\cassandra\data\system\LocationInfo-e-1-Data.db at org.apache.cassandra.io.sstable.SSTableWriter.rename(SSTableWriter.java: 214) at org.apache.cassandra.io.sstable.SSTableWriter.closeAndOpenReader(SSTable Writer.java:184) at org.apache.cassandra.io.sstable.SSTableWriter.closeAndOpenReader(SSTable Writer.java:167) at org.apache.cassandra.db.Memtable.writeSortedContents(Memtable.java:161) at org.apache.cassandra.db.Memtable.access$000(Memtable.java:49) at org.apache.cassandra.db.Memtable$1.runMayThrow(Memtable.java:174) at org.apache.cassandra.utils.WrappedRunnable.run(WrappedRunnable.java:30) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:441) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecuto r.java:886) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.ja va:908) at java.lang.Thread.run(Thread.java:619) Caused by: java.io.IOException: rename failed of d:\cassandra\data\system\LocationInfo-e-1-Data.db at org.apache.cassandra.utils.FBUtilities.renameWithConfirm(FBUtilities.jav a:359) at org.apache.cassandra.io.sstable.SSTableWriter.rename(SSTableWriter.java: 210) ... 12 more Operating system is Windows 7. Tried it also on Windows 2003 server. I only modified a few (necessary) path settings in cassandra.yaml: commitlog_directory: d:/cassandra/commitlog data_file_directories: - d:/cassandra/data saved_caches_directory: d:/cassandra/saved_caches Does anybody know what I'm doing wrong? Regards, Ramon -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of Riptano, the source for professional Cassandra support http://riptano.com Geen virus gevonden in dit bericht. Gecontroleerd door AVG - www.avg.com Versie: 10.0.1170 / Virusdatabase: 426/3286 - datum van uitgifte: 11/28/10
Re: Introduction to Cassandra
Jim, Jonathan thanks for the feedback. I was trying something different and rattled through the whole deck in about 35-40 minutes. I'll be doing the talk again at the Wellington Python Users Group this Thursday http://nzpug.org/MeetingsWellington Aaron On 30 Nov 2010, at 20:18, Jim Morrison wrote: Really great introduction, thanks Aaron.Bookmarked for the team. J. Sent from my iPhone On 29 Nov 2010, at 21:11, Aaron Morton aa...@thelastpickle.com wrote: I did a talk last week at the Wellington Rails User Group as a basic introduction to Cassandra. The slides are here http://www.slideshare.net/aaronmorton/well-railedcassandra24112010-5901169 if anyone is interested. Cheers Aaron
Re: Achieving isolation on single row modifications with batch_mutate
I'm chunking up a larger blob. Basically the size of each row can vary (averages around 500K - 1MB), with some outliers in the 50 MB range. However, when I do an update, I can usually just read/update a portion of that blob. A lot of my read operations can also work on a smaller chunk. The number of columns is going to depend on the size of the blob itself. I'm also considering using supercolumns to have higher save granularity. My biggest problem is that I will have to update these rows a lot (several times a day) and often very quickly (process 15 thousand in 2-3 minutes). While I think I could probably scale up with a lot of hardware to meet that load, it seems like I'm doing much much more work than I need to (processing 15 GB of data in 2-3 minutes as opposes to 100 MB). I also worry about handling our future data size needs. I can split the blob up without a lot of extra complexity but am worried about how to have readers read a non-corrupted version of the object, since sometimes I'll have to update multiple chunks as one unit. From: Tyler Hobbs ty...@riptano.com To: user@cassandra.apache.org Sent: Tue, November 30, 2010 12:57:07 AM Subject: Re: Achieving isolation on single row modifications with batch_mutate In this case, it sounds like you should combine columns A and B if you are writing them both at the same time, reading them both at the same time, and need them to be consistent. Obviously, you're probably dealing with more than two columns here, but there's generally not any value in splitting something into multiple columns if you're always writing and reading all of them at the same time. Or are you talking about chunking huge blobs across a row? - Tyler On Sat, Nov 27, 2010 at 10:12 AM, E S tr1skl...@yahoo.com wrote: I'm trying to figure out the best way to achieve single row modification isolation for readers. As an example, I have 2 rows (1,2) with 2 columns (a,b). If I modify both rows, I don't care if the user sees the write operations completed on 1 and not on 2 for a short time period (seconds). I also don't care if when reading row 1 the user gets the new value, and then on a re-read gets the old value (within a few seconds). Because of this, I have been planning on using a consistency level of one. However, if I modify both columns A,B on a single row, I need both changes on the row to be visible/invisible atomically. It doesn't matter if they both become visible and then both invisible as the data propagates across nodes, but a half-completed state on an initial read will basically be returning corrupt data given my apps consistency requirements. My understanding from the FAQ that this single row multicolumn change provides no read isolation, so I will have this problem. Is this correct? If so: Question 1: Is there a way to get this type of isolation without using a distributed locking mechanism like cages? Question 2: Are there any plans to implement this type of isolation within Cassandra? Question 3: If I went with a distributed locking mechanism, what consistency level would I need to use with Cassandra? Could I still get away with a consistency level of one? It seems that if the initial write is done in a non-isolated way, but if cross-node row synchronizations are done all or nothing, I could still use one. Question 4: Does anyone know of a good c# alternative to cages/zookeeper? Thanks for any help with this!
Re: Updating Cascal
Hi Tyler, Thanks for the response. I decided to give up on it, and start my own Scala based api modeled on Cascal since it's no longer supported. _M!ke On Nov 30, 2010, at 1:06 AM, Tyler Hobbs wrote: Are you sure you're using the same key for batch_mutate() and get_slice()? They appear different in the logs. - Tyler On Thu, Nov 25, 2010 at 10:14 AM, Michael Fortin mi...@m410.us wrote: Hello, I forked Cascal (Scala based client for cassandra) and I'm attempting to update it to cassandra 0.7. I have it partially working, but I'm getting stuck on a few areas. I have most of the unit tests working from the original code, but I'm having an issue with batch_mutate(keyToFamilyMutations, consistency) . Does the log output mean anything? I can't figure out why the columns are not getting inserted. If I change th code from a batch_mutate to an insert(family, parent, column, consistency) it works. ### keyToFamilyMutations: {java.nio.HeapByteBuffer[pos=0 lim=16 cap=16]={Standard=[Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 31, value:56 61 6C 75 65 2D 31, timestamp:1290662894466035))), Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 33, value:56 61 6C 75 65 2D 33, timestamp:1290662894467942))), Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 32, value:56 61 6C 75 65 2D 32, timestamp:1290662894467915)))]}} DEBUG 2010-11-25 00:28:14,534 [org.apache.cassandra.thrift.CassandraServer pool-1-thread-2] batch_mutate DEBUG 2010-11-25 00:28:14,583 [org.apache.cassandra.service.StorageProxy pool-1-thread-2] insert writing local RowMutation(keyspace='Test', key='ccfd5520f85411df858a001c4209', modifications=[Standard]) DEBUG 2010-11-25 00:28:14,599 [org.apache.cassandra.thrift.CassandraServer pool-1-thread-2] get_slice DEBUG 2010-11-25 00:28:14,605 [org.apache.cassandra.service.StorageProxy pool-1-thread-2] weakread reading SliceFromReadCommand(table='Test', key='5374616e64617264', column_parent='QueryPath(columnFamilyName='Standard', superColumnName='null', columnName='null')', start='', finish='', reversed=false, count=2147483647) locally DEBUG 2010-11-25 00:28:14,608 [org.apache.cassandra.service.StorageProxy ReadStage:2] weakreadlocal reading SliceFromReadCommand(table='Test', key='5374616e64617264', column_parent='QueryPath(columnFamilyName='Standard', superColumnName='null', columnName='null')', start='', finish='', reversed=false, count=2147483647) ### get_slice: [] The code looks like: println(keyToFamilyMutations: %s.format(keyToFamilyMutations)) client.batch_mutate(keyToFamilyMutations, consistency) … client.client.get_slice(…) keyspaces: - name: Test replica_placement_strategy: org.apache.cassandra.locator.SimpleStrategy replication_factor: 1 column_families: - {name: Standard, compare_with: BytesType} Thanks, Mike
Re: Achieving isolation on single row modifications with batch_mutate
On Sat, Nov 27, 2010 at 10:12 AM, E S tr1skl...@yahoo.com wrote: I'm trying to figure out the best way to achieve single row modification isolation for readers. I have a lot of No's for you. :) As an example, I have 2 rows (1,2) with 2 columns (a,b). If I modify both rows, I don't care if the user sees the write operations completed on 1 and not on 2 for a short time period (seconds). I also don't care if when reading row 1 the user gets the new value, and then on a re-read gets the old value (within a few seconds). Because of this, I have been planning on using a consistency level of one. However, if I modify both columns A,B on a single row, I need both changes on the row to be visible/invisible atomically. It doesn't matter if they both become visible and then both invisible as the data propagates across nodes, but a half-completed state on an initial read will basically be returning corrupt data given my apps consistency requirements. My understanding from the FAQ that this single row multicolumn change provides no read isolation, so I will have this problem. Is this correct? If so: Question 1: Is there a way to get this type of isolation without using a distributed locking mechanism like cages? No. Question 2: Are there any plans to implement this type of isolation within Cassandra? No. Question 3: If I went with a distributed locking mechanism, what consistency level would I need to use with Cassandra? Could I still get away with a consistency level of one? Maybe. If you want to guarantee that you see the most recent write, then ONE will not be high enough. But if all you care about is seeing all of the update or none of it, then ONE + locking will be fine. Question 4: Does anyone know of a good c# alternative to cages/zookeeper? No. -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of Riptano, the source for professional Cassandra support http://riptano.com
Re: Updating Cascal
Did you look at Scromium? On Tue, Nov 30, 2010 at 8:27 AM, Michael Fortin mi...@m410.us wrote: Hi Tyler, Thanks for the response. I decided to give up on it, and start my own Scala based api modeled on Cascal since it's no longer supported. _M!ke On Nov 30, 2010, at 1:06 AM, Tyler Hobbs wrote: Are you sure you're using the same key for batch_mutate() and get_slice()? They appear different in the logs. - Tyler On Thu, Nov 25, 2010 at 10:14 AM, Michael Fortin mi...@m410.us wrote: Hello, I forked Cascal (Scala based client for cassandra) and I'm attempting to update it to cassandra 0.7. I have it partially working, but I'm getting stuck on a few areas. I have most of the unit tests working from the original code, but I'm having an issue with batch_mutate(keyToFamilyMutations, consistency) . Does the log output mean anything? I can't figure out why the columns are not getting inserted. If I change th code from a batch_mutate to an insert(family, parent, column, consistency) it works. ### keyToFamilyMutations: {java.nio.HeapByteBuffer[pos=0 lim=16 cap=16]={Standard=[Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 31, value:56 61 6C 75 65 2D 31, timestamp:1290662894466035))), Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 33, value:56 61 6C 75 65 2D 33, timestamp:1290662894467942))), Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 32, value:56 61 6C 75 65 2D 32, timestamp:1290662894467915)))]}} DEBUG 2010-11-25 00:28:14,534 [org.apache.cassandra.thrift.CassandraServer pool-1-thread-2] batch_mutate DEBUG 2010-11-25 00:28:14,583 [org.apache.cassandra.service.StorageProxy pool-1-thread-2] insert writing local RowMutation(keyspace='Test', key='ccfd5520f85411df858a001c4209', modifications=[Standard]) DEBUG 2010-11-25 00:28:14,599 [org.apache.cassandra.thrift.CassandraServer pool-1-thread-2] get_slice DEBUG 2010-11-25 00:28:14,605 [org.apache.cassandra.service.StorageProxy pool-1-thread-2] weakread reading SliceFromReadCommand(table='Test', key='5374616e64617264', column_parent='QueryPath(columnFamilyName='Standard', superColumnName='null', columnName='null')', start='', finish='', reversed=false, count=2147483647) locally DEBUG 2010-11-25 00:28:14,608 [org.apache.cassandra.service.StorageProxy ReadStage:2] weakreadlocal reading SliceFromReadCommand(table='Test', key='5374616e64617264', column_parent='QueryPath(columnFamilyName='Standard', superColumnName='null', columnName='null')', start='', finish='', reversed=false, count=2147483647) ### get_slice: [] The code looks like: println(keyToFamilyMutations: %s.format(keyToFamilyMutations)) client.batch_mutate(keyToFamilyMutations, consistency) … client.client.get_slice(…) keyspaces: - name: Test replica_placement_strategy: org.apache.cassandra.locator.SimpleStrategy replication_factor: 1 column_families: - {name: Standard, compare_with: BytesType} Thanks, Mike -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of Riptano, the source for professional Cassandra support http://riptano.com
Re: Updating Cascal
Your referring to this: https://github.com/cliffmoon/scromium right? Thanks for the tip, I'll give it a try. _Mike On Nov 30, 2010, at 9:51 AM, Jonathan Ellis wrote: Did you look at Scromium? On Tue, Nov 30, 2010 at 8:27 AM, Michael Fortin mi...@m410.us wrote: Hi Tyler, Thanks for the response. I decided to give up on it, and start my own Scala based api modeled on Cascal since it's no longer supported. _M!ke On Nov 30, 2010, at 1:06 AM, Tyler Hobbs wrote: Are you sure you're using the same key for batch_mutate() and get_slice()? They appear different in the logs. - Tyler On Thu, Nov 25, 2010 at 10:14 AM, Michael Fortin mi...@m410.us wrote: Hello, I forked Cascal (Scala based client for cassandra) and I'm attempting to update it to cassandra 0.7. I have it partially working, but I'm getting stuck on a few areas. I have most of the unit tests working from the original code, but I'm having an issue with batch_mutate(keyToFamilyMutations, consistency) . Does the log output mean anything? I can't figure out why the columns are not getting inserted. If I change th code from a batch_mutate to an insert(family, parent, column, consistency) it works. ### keyToFamilyMutations: {java.nio.HeapByteBuffer[pos=0 lim=16 cap=16]={Standard=[Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 31, value:56 61 6C 75 65 2D 31, timestamp:1290662894466035))), Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 33, value:56 61 6C 75 65 2D 33, timestamp:1290662894467942))), Mutation(column_or_supercolumn:ColumnOrSuperColumn(column:Column(name:43 6F 6C 75 6D 6E 2D 61 2D 32, value:56 61 6C 75 65 2D 32, timestamp:1290662894467915)))]}} DEBUG 2010-11-25 00:28:14,534 [org.apache.cassandra.thrift.CassandraServer pool-1-thread-2] batch_mutate DEBUG 2010-11-25 00:28:14,583 [org.apache.cassandra.service.StorageProxy pool-1-thread-2] insert writing local RowMutation(keyspace='Test', key='ccfd5520f85411df858a001c4209', modifications=[Standard]) DEBUG 2010-11-25 00:28:14,599 [org.apache.cassandra.thrift.CassandraServer pool-1-thread-2] get_slice DEBUG 2010-11-25 00:28:14,605 [org.apache.cassandra.service.StorageProxy pool-1-thread-2] weakread reading SliceFromReadCommand(table='Test', key='5374616e64617264', column_parent='QueryPath(columnFamilyName='Standard', superColumnName='null', columnName='null')', start='', finish='', reversed=false, count=2147483647) locally DEBUG 2010-11-25 00:28:14,608 [org.apache.cassandra.service.StorageProxy ReadStage:2] weakreadlocal reading SliceFromReadCommand(table='Test', key='5374616e64617264', column_parent='QueryPath(columnFamilyName='Standard', superColumnName='null', columnName='null')', start='', finish='', reversed=false, count=2147483647) ### get_slice: [] The code looks like: println(keyToFamilyMutations: %s.format(keyToFamilyMutations)) client.batch_mutate(keyToFamilyMutations, consistency) … client.client.get_slice(…) keyspaces: - name: Test replica_placement_strategy: org.apache.cassandra.locator.SimpleStrategy replication_factor: 1 column_families: - {name: Standard, compare_with: BytesType} Thanks, Mike -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of Riptano, the source for professional Cassandra support http://riptano.com
Re: Updating Cascal
I'd highly recommend looking at Hector (v2) as well. It's very nice. I'm using it from Scala without any issues. Rather than duplicating the effort of scromium, cascal, scalandra, and not to mention Hector itself, perhaps it'd worthwhile taking a stab at a Scala interface wrapping Hector? Connection/failover strategies, test helpers and client metrics aren't THAT much fun reinventing. /d On Nov 30, 2010, at 9:51 AM, Jonathan Ellis wrote: Did you look at Scromium? On Tue, Nov 30, 2010 at 8:27 AM, Michael Fortin mi...@m410.us wrote: Hi Tyler, Thanks for the response. I decided to give up on it, and start my own Scala based api modeled on Cascal since it's no longer supported.
Re: Achieving isolation on single row modifications with batch_mutate
It's hard to tell without knowing the the nature of the data you're writing, but you might want to think about whether you can embed any sort of version number and/or checksum into the column names of the chunk columns. That way, you could very easily determine that the data you wanted to retrieve was not yet available for reading. Are you able do your partial blob updates on an entire chunk at a time or do you need to read the blob chunk, modify a portion of it, and then write it back? If it's the former, then it might be possible for this to be accomplished without a locking solution. Ed On Sat, Nov 27, 2010 at 8:12 AM, E S tr1skl...@yahoo.com wrote: I'm trying to figure out the best way to achieve single row modification isolation for readers. As an example, I have 2 rows (1,2) with 2 columns (a,b). If I modify both rows, I don't care if the user sees the write operations completed on 1 and not on 2 for a short time period (seconds). I also don't care if when reading row 1 the user gets the new value, and then on a re-read gets the old value (within a few seconds). Because of this, I have been planning on using a consistency level of one. However, if I modify both columns A,B on a single row, I need both changes on the row to be visible/invisible atomically. It doesn't matter if they both become visible and then both invisible as the data propagates across nodes, but a half-completed state on an initial read will basically be returning corrupt data given my apps consistency requirements. My understanding from the FAQ that this single row multicolumn change provides no read isolation, so I will have this problem. Is this correct? If so: Question 1: Is there a way to get this type of isolation without using a distributed locking mechanism like cages? Question 2: Are there any plans to implement this type of isolation within Cassandra? Question 3: If I went with a distributed locking mechanism, what consistency level would I need to use with Cassandra? Could I still get away with a consistency level of one? It seems that if the initial write is done in a non-isolated way, but if cross-node row synchronizations are done all or nothing, I could still use one. Question 4: Does anyone know of a good c# alternative to cages/zookeeper? Thanks for any help with this!
Re: get_count - cassandra 0.7.x predicate limit bug?
On Tue, Nov 30, 2010 at 1:00 AM, Tyler Hobbs ty...@riptano.com wrote: What error are you getting? Remember, get_count() is still just about as much work for cassandra as getting the whole row; the only advantage is it doesn't have to send the whole row back to the client. If you're counting 3+ million columns frequently, it's time to take a look at counters. - Tyler On Fri, Nov 26, 2010 at 10:33 AM, Marcin mar...@33concept.com wrote: Hi guys, I have a key with 3million+ columns but when I am trying to run get_count on it its getting me error if setting limit more than 46000+ any ideas? In previous API there was no predicate at all so it was simply counting number of columns now its not so simple any more. Please let me know if that is a bug or I do something wrong. cheers, /Marcin +1 Tyler. The problem is you can increase the clients socket timeout as high as you like if socketTimeout rpcTimeout you should see SocketTimeoutExceptions if socketTimeout = rcpTimeout you start seeing Cassandra TimedOutExceptions. Raising the RPC Timeout is done on the server. In any case you may have to range_slice to get through a row this big and count. Also in my experience rows this large do not work well. They are particularly dangerous when combined with RowCache as bringing them into to memory and evicting them is both disk and memory intensive.
JVM OOM on node startup
Hello again. We have 3 nodes and were testing what happens when a node goes down. There is roughly 10gb of data on each node. The node we simulated dieing was working just fine under the load. Then we killed it. The ring performed admirably, But upon restarting the node it dies every time of JVM OOM errors. I have forced a JVM heap size of 1024mb in the startup file. (did this because adaptive heap size was causing oom errors with normal usage.) The machines are 2 core 4gb ram vm's. I've read the Riptano troubleshooting guide... http://www.riptano.com/docs/0.6/troubleshooting/index#nodes-are-dying-with-oom-errors But im not sure if these apply in this case since it is only dieing on startup. Here is a link to the startup logs as it dies. http://pastebin.com/BEXeVvCX Thank you for any help you can provide.
Re: JVM OOM on node startup
Looks like it's trying to load your row cache and running out of memory, probably because you reduced the memory. The cassandra-env.sh script would have been giving it 2GB.1Gb heap is probably going to be to small.Was this the same error you were getting before you reduced the memory ?Try deleting the caches, the path is specified by the saved_caches_directory setting in cassandra.yaml.Also what version are you using ? The errorCaused by: javax.management.AttributeNotFoundException: No such attribute: ActiveCount reminds me of a problem in beta 1.Hope that helps.AaronOn 01 Dec, 2010,at 09:28 AM, Brayton Thompson thomp...@grnoc.iu.edu wrote:Hello again. We have 3 nodes and were testing what happens when a node goes down. There is roughly 10gb of data on each node. The node we "simulated" dieing was working just fine under the load. Then we killed it. The ring performed admirably, But upon restarting the node it dies every time of JVM OOM errors. I have forced a JVM heap size of 1024mb in the startup file. (did this because adaptive heap size was causing oom errors with normal usage.) The machines are 2 core 4gb ram vm's. I've read the Riptano troubleshooting guide... http://www.riptano.com/docs/0.6/troubleshooting/index#nodes-are-dying-with-oom-errors But im not sure if these apply in this case since it is only dieing on startup. Here is a link to the startup logs as it dies. http://pastebin.com/BEXeVvCX Thank you for any help you can provide.
Re: JVM OOM on node startup
If you're getting OOM with adaptive heap size of 1GB, reducing it to 1GB is not going to make things better. :) On Tue, Nov 30, 2010 at 2:28 PM, Brayton Thompson thomp...@grnoc.iu.edu wrote: Hello again. We have 3 nodes and were testing what happens when a node goes down. There is roughly 10gb of data on each node. The node we simulated dieing was working just fine under the load. Then we killed it. The ring performed admirably, But upon restarting the node it dies every time of JVM OOM errors. I have forced a JVM heap size of 1024mb in the startup file. (did this because adaptive heap size was causing oom errors with normal usage.) The machines are 2 core 4gb ram vm's. I've read the Riptano troubleshooting guide... http://www.riptano.com/docs/0.6/troubleshooting/index#nodes-are-dying-with-oom-errors But im not sure if these apply in this case since it is only dieing on startup. Here is a link to the startup logs as it dies. http://pastebin.com/BEXeVvCX Thank you for any help you can provide. -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of Riptano, the source for professional Cassandra support http://riptano.com
Re: Achieving isolation on single row modifications with batch_mutate
I'm a little confused about #3. Hopefully this clarifying question won't turn the one maybe into a no :). I'm fine not reading the latest data, as long as on each individual read I see all or none of the operations that occurred for a single one row batch_mutate. My concern is do I have to lock the reads until they have propagated to all nodes. If I do a batch_mutate with a consistency of ONE onto one row, during the write operation to the one node a reader can see partial changes. Once the batch mutate completes, the change has not been propagated to the other nodes. On a per row basis, are the changes to other nodes pushed in an isolated manner? If not, it seems like I would have to write with a consistency of ALL and lock around that. - Original Message From: Jonathan Ellis jbel...@gmail.com To: user user@cassandra.apache.org Sent: Tue, November 30, 2010 9:50:51 AM Subject: Re: Achieving isolation on single row modifications with batch_mutate On Sat, Nov 27, 2010 at 10:12 AM, E S tr1skl...@yahoo.com wrote: I'm trying to figure out the best way to achieve single row modification isolation for readers. I have a lot of No's for you. :) As an example, I have 2 rows (1,2) with 2 columns (a,b). If I modify both rows, I don't care if the user sees the write operations completed on 1 and not on 2 for a short time period (seconds). I also don't care if when reading row 1 the user gets the new value, and then on a re-read gets the old value (within a few seconds). Because of this, I have been planning on using a consistency level of one. However, if I modify both columns A,B on a single row, I need both changes on the row to be visible/invisible atomically. It doesn't matter if they both become visible and then both invisible as the data propagates across nodes, but a half-completed state on an initial read will basically be returning corrupt data given my apps consistency requirements. My understanding from the FAQ that this single row multicolumn change provides no read isolation, so I will have this problem. Is this correct? If so: Question 1: Is there a way to get this type of isolation without using a distributed locking mechanism like cages? No. Question 2: Are there any plans to implement this type of isolation within Cassandra? No. Question 3: If I went with a distributed locking mechanism, what consistency level would I need to use with Cassandra? Could I still get away with a consistency level of one? Maybe. If you want to guarantee that you see the most recent write, then ONE will not be high enough. But if all you care about is seeing all of the update or none of it, then ONE + locking will be fine. Question 4: Does anyone know of a good c# alternative to cages/zookeeper? No. -- Jonathan Ellis Project Chair, Apache Cassandra co-founder of Riptano, the source for professional Cassandra support http://riptano.com
pid file is not created in Windows Environment....
Hi, Can you Please help me, why the pid file is not created in windows environment when I try with C:\apache-cassandra-0.6.6\bincassandra.bat -p c.pid ? Is there a better way to shutdown the cassandra server instead of kill pid? Thanks, -Ram.
C++ client for Cassandra
Are there any C++ clients out there similar to Hector (in terms of features) for Cassandra? I am looking for C++ Client for Cassandra 0.7. Thanks, Naren
Re: C++ client for Cassandra
Thrift is there.. On Wed, Dec 1, 2010 at 11:43 AM, Narendra Sharma narendra.sha...@gmail.comwrote: Are there any C++ clients out there similar to Hector (in terms of features) for Cassandra? I am looking for C++ Client for Cassandra 0.7. Thanks, Naren
When to call the major compaction ?
Every time cassandra creates a new sstable , it will call the CompactionManager.submitMinorIfNeeded ? And if the number of memtables is beyond MinimumCompactionThreshold , the minor compaction will be called. And there is also a method named CompactionManager.submitMajor , and the call relationship is : NodeCmd -- NodeProbe --StorageService.forceTableCompaction -- Table.forceCompaction --CompactionManager.performMajor -- CompactionManager.submitMajor ColumnFamilyStore.forceMajorCompaction -- CompactionManager.performMajor -- CompactionManager.submitMajor HintedHandOffManager -- CompactionManager.submitMajor So i have 3 questions: 1. Once a new sstable has been created , CompactionManager.submitMinorIfNeeded will be called , minorCompaction maybe called . But when will the majorCompaction be called ? Just the NodeCmd ? 2. Which jobs will minorCompaction and majorCompaction do ? Will minorCompaction delete the data that have been marked as deleted ? And how about the major compaction ? 3. When gc be called ? Every time compaction been called? -- Best regards, Ivy Tang