2018-03-21 09:11:20 UTC - Piotr: @Ivan Kelly thanks. Didn´t see that in the official docs (<https://pulsar.incubator.apache.org/docs/latest/clients/Java/>) ---- 2018-03-21 09:12:00 UTC - Ivan Kelly: Indeed. A ghastly oversight. Will fix ---- 2018-03-21 09:13:26 UTC - Ivan Kelly: had you clicked through to <https://pulsar.incubator.apache.org/docs/latest/admin-api/overview/>? ---- 2018-03-21 09:13:42 UTC - Ivan Kelly: trying to decide where it should be fixed ---- 2018-03-21 09:14:08 UTC - Piotr: Hehe, no worries. I did yes. I think it makes sense to add it to the first one ---- 2018-03-21 09:15:13 UTC - Piotr: and perhaps just link to the first one from the second one? ---- 2018-03-21 09:17:18 UTC - Ivan Kelly: ok. will look into it ---- 2018-03-21 09:25:39 UTC - Piotr: Also, do you know of anyone running managed pulsar clusters yet? (Saas). I saw that streamlio seem to be going for it but don´t have a public offering yet. ---- 2018-03-21 09:31:35 UTC - Ivan Kelly: <https://github.com/apache/incubator-pulsar/pull/1407> ---- 2018-03-21 09:31:49 UTC - Ivan Kelly: Noone offering as Saas yet afaik ---- 2018-03-21 10:26:31 UTC - Piotr: @Ivan Kelly seems like the pulsar-client-admin dep is not enough? Is it missing some deps, or have I missed something again? :slightly_smiling_face: Failed to instantiate [org.apache.pulsar.client.admin.PulsarAdmin]: Factory method 'pulsarAdmin' threw exception; nested exception is java.lang.RuntimeException: java.lang.ClassNotFoundException: Provider org.apache.pulsar.admin.shade.org.glassfish.jersey.internal.RuntimeDelegateImpl could not be instantiated: java.lang.IllegalStateException: No generator was provided and there is no default generator registered ---- 2018-03-21 10:27:06 UTC - Ivan Kelly: have you tried pulling in the client also? ---- 2018-03-21 10:27:29 UTC - Piotr: yes I have both in my pom ---- 2018-03-21 10:29:22 UTC - Piotr: also tried adding jersey-common after some googling but that didn´t help ---- 2018-03-21 10:30:27 UTC - Ivan Kelly: ok. there's a bug there ---- 2018-03-21 10:30:41 UTC - Ivan Kelly: it's a shading issue. Could you create an issue on github? ---- 2018-03-21 10:31:04 UTC - Piotr: Will do ---- 2018-03-21 10:31:16 UTC - Ivan Kelly: try pulling in pulsar-client-admin-original instead of pulsar-client-admin ---- 2018-03-21 10:31:45 UTC - Ivan Kelly: it doesn't shade the dependencies, so this issue goes away though you may be clashes on other stuff ---- 2018-03-21 10:37:10 UTC - Piotr: That worked, and issue submitted. Thanks! ---- 2018-03-21 10:37:34 UTC - Ivan Kelly: np ---- 2018-03-21 10:37:44 UTC - Piotr: Super excited about using pulsar. I was just about to give up my search and go for a less ideal solution with rabbitmq when I found it ---- 2018-03-21 10:39:07 UTC - Ivan Kelly: great to hear :slightly_smiling_face: let me know if you have any questions. ---- 2018-03-21 10:43:02 UTC - Piotr: thanks. Hopefully I won´t have too many annoying ones :grimacing: ---- 2018-03-21 12:19:17 UTC - Piotr: Another head scratcher. Seems like some jackson version problem: ---- 2018-03-21 12:19:19 UTC - Piotr: java.lang.NoSuchMethodError: com.fasterxml.jackson.jaxrs.cfg.EndpointConfigBase.<init>(Lcom/fasterxml/jackson/databind/cfg/MapperConfig;)V ---- 2018-03-21 12:21:42 UTC - Piotr: Which doesn´t make sense as that constructor does exist in the jackson version provided by the parent POM (<https://github.com/apache/incubator-pulsar/blob/v1.22.0-incubating/pom.xml>) ---- 2018-03-21 12:22:01 UTC - Piotr: Any ideas? ---- 2018-03-21 12:56:48 UTC - Piotr: It seems that for some reason the version of jackson-jaxrs-base is different (2.3.2) from the other jackson deps (2.9.3). Trying to figure out why ---- 2018-03-21 13:14:54 UTC - Piotr: Ok I think I figured out why. It seems that the whole pulsar project depends on jersey-media-json-jackson 2.9, which in turn depends on jackson-jaxrs-base 2.3.2 while the jackson version used in the pulsar project is actually much higher (2.9.3). Is this a bug or is it just because I use the pulsar-client-admin artifact? ---- 2018-03-21 13:29:29 UTC - Piotr: I was able to work around it by excluding the old version and including the new one of jackson-jaxrs-json-provider. And commented about this in the github issue. ---- 2018-03-21 13:41:36 UTC - Ivan Kelly: ya, these are the other conflicts I mentioned. ---- 2018-03-21 13:42:38 UTC - Piotr: alright. Well, got it working for now at least. On to the fun stuff ---- 2018-03-21 14:00:43 UTC - Piotr: Ok, next question :sweat_smile:I get a cryptic error message when I do this: "pulsar-admin persistent lookup <persistent://myapp/global/entity-updates/test-entity-id>". It just gives me "HTTP 412 Precondition Failed". And I can not find documentation of the /lookup/ resource in the docs to explain it (<https://pulsar.incubator.apache.org/docs/latest/reference/RestApi/>) ---- 2018-03-21 14:01:22 UTC - Piotr: (I´m trying to check if a topic exists, so that I can create it if it doesn´t) ---- 2018-03-21 14:02:18 UTC - Piotr: (Reason why I am attracted by pulsar instead of kafka is because kafka was not built to handle a massive amount of topics, but pulsar should be able to right?) ---- 2018-03-21 14:43:56 UTC - Ivan Kelly: is that the full error? just precondition failed? ---- 2018-03-21 14:44:18 UTC - Ivan Kelly: pulsar can take loads of topics, yes ---- 2018-03-21 14:44:19 UTC - Piotr: Unfortunately yes ---- 2018-03-21 14:44:29 UTC - Ivan Kelly: ok, let me check the souce ---- 2018-03-21 14:44:35 UTC - Piotr: Maybe there is some log I can check for the exact cause? ---- 2018-03-21 14:44:45 UTC - Ivan Kelly: how are you running pulsar? ---- 2018-03-21 14:44:57 UTC - Piotr: I installed it using streamlio:s homebrew recipe ---- 2018-03-21 14:48:32 UTC - Ivan Kelly: not sure where logs are with that ---- 2018-03-21 14:48:43 UTC - Ivan Kelly: don't use a Mac myself ---- 2018-03-21 14:49:18 UTC - Piotr: ok I am sure I can find it myself, do you know the log file name? ---- 2018-03-21 14:50:01 UTC - Ivan Kelly: broker.log ---- 2018-03-21 14:51:41 UTC - Ivan Kelly: what files are there with .log under the top level directory? ---- 2018-03-21 14:54:23 UTC - Piotr: hmm found the binaries in /usr/local/bin but still looking for the logs. Not in /var/logs . Any other ideas? ---- 2018-03-21 14:55:40 UTC - Ivan Kelly: jps ---- 2018-03-21 14:55:44 UTC - Ivan Kelly: find process ---- 2018-03-21 14:55:49 UTC - Ivan Kelly: lsof -p <pid> ---- 2018-03-21 14:56:50 UTC - Piotr: alright found it ---- 2018-03-21 14:56:54 UTC - Piotr: thanks ---- 2018-03-21 14:57:36 UTC - Piotr: Perfect. Got the error right there (2018-03-21 15:57:08,843 - ERROR - [pulsar-web-76-4:DestinationLookup@99] - Validation check failed: Global namespace does not have any clusters configured) ---- 2018-03-21 14:57:51 UTC - Piotr: I guess it´s a bug that the error message is not emitted to the client ---- 2018-03-21 14:58:01 UTC - Ivan Kelly: ah ---- 2018-03-21 14:58:03 UTC - Piotr: Want me to file a bug? :smile: ---- 2018-03-21 14:58:11 UTC - Ivan Kelly: please do ---- 2018-03-21 14:58:32 UTC - Ivan Kelly: you need to set-allowed-clusters on the global namespace ---- 2018-03-21 14:58:45 UTC - Piotr: got it, thanks ---- 2018-03-21 15:14:38 UTC - Piotr: Another bug? I try to update my property to allow the global cluster. But I can´t get the update command to work. "pulsar-admin properties update myapp -c global" and "pulsar-admin properties update -c global myapp" both return the Usage page. ---- 2018-03-21 15:16:01 UTC - Matteo Merli: There are 2 required arguments to the "update" command ---- 2018-03-21 15:16:28 UTC - Piotr: ah there you go. Thanks ---- 2018-03-21 15:16:44 UTC - Matteo Merli: Also, you need to specify the actual cluster names rather than "global" there ---- 2018-03-21 15:18:49 UTC - Piotr: really? It did go through but I do get this in the log now: "Global namespace does not have any clusters configured". Is that why? ---- 2018-03-21 15:19:16 UTC - Piotr: Am I not supposed to use the global cluster directly? ---- 2018-03-21 15:21:02 UTC - Matteo Merli: Yes, it's a part that's a bit confusing. We're working on sinplyfyng the terminology and the context there. The global name, means the namespace can be replicated across multiple clusters ---- 2018-03-21 15:21:26 UTC - Matteo Merli: Then you configure the list of replication clusters for a namespace. ---- 2018-03-21 15:22:07 UTC - Matteo Merli: For a property (tenant), the allowed clusters are the "real" clusters that this tenant can use ---- 2018-03-21 15:22:56 UTC - Piotr: ok, so I can´t just use "global" as a wildcard? ---- 2018-03-21 15:23:01 UTC - Matteo Merli: Even if you have 3 clusters, (a, b and c), a particular tenant might be restricted to just 2 of them ---- 2018-03-21 15:23:12 UTC - Piotr: Or, where can/should I use it etc if I want the topics to be available in all clusters? :stuck_out_tongue: ---- 2018-03-21 15:23:49 UTC - Matteo Merli: Unfortunately not, you have to allow it on all clusters, with the list ---- 2018-03-21 15:24:12 UTC - Matteo Merli: We're in process of simplifying that part as well ---- 2018-03-21 15:24:50 UTC - Piotr: ok no worries. So even in the app code I have to use a specific cluster name? So do I use the global name anywhere? ---- 2018-03-21 15:26:08 UTC - Piotr: pulsar-admin namespaces set-clusters -c standalone myapp/global/entity-updates then resulted in ---- 2018-03-21 15:26:24 UTC - Piotr: pulsar-admin namespaces get-clusters myapp/global/entity-updates = standalone ---- 2018-03-21 15:26:33 UTC - Piotr: so I can use global for querying, but not for setting? ---- 2018-03-21 15:27:55 UTC - Piotr: (if you have any docs about this I can stop bugging you and read that instead) ---- 2018-03-21 16:04:04 UTC - Matteo Merli: @Piotr take a look at <http://pulsar.apache.org/docs/latest/admin/GeoReplication/#Configuringreplication-rs8in4> ---- 2018-03-21 16:04:18 UTC - Matteo Merli: that has an example ---- 2018-03-21 16:04:50 UTC - Matteo Merli: basically you need to : * Create a property (tenant) that is allowed to use the specific clusters ---- 2018-03-21 16:05:47 UTC - Matteo Merli: * Create a “global” namespace. eg: `my-property/global/my-namespace` * Configure replication for that namespace ```bin/pulsar-admin namespaces set-clusters my-property/global/my-namespace \ --clusters us-west,us-east,us-cent ``` ---- 2018-03-21 16:10:36 UTC - Deepak Bansal: @Deepak Bansal has joined the channel ---- 2018-03-21 19:02:18 UTC - Piotr: thanks, that cleared it up a bit more. ---- 2018-03-21 19:03:04 UTC - Piotr: Is there any way to prevent topic garbage collection? What if I have infinite message retention? The docs say "a topic is automatically deleted when it’s no longer used, that is to say, when no producers or consumers are connected and there are no subscriptions." ---- 2018-03-21 19:04:03 UTC - Matteo Merli: so, if there is “time retention”, the topic is also not deleted until that expires (which can be never if retention is -1 (infinite)) ---- 2018-03-21 19:04:33 UTC - Piotr: Ah ok - want me to file an issue to change the docs? :wink: I think it is rather important to know ---- 2018-03-21 19:05:39 UTC - Matteo Merli: :+1: ---- 2018-03-21 19:05:51 UTC - Matteo Merli: thanks ---- 2018-03-21 19:11:32 UTC - Matteo Merli: there’s also the option to completely turn off topic auto-deletion: `brokerDeleteInactiveTopicsEnabled=false` ---- 2018-03-21 19:13:19 UTC - Piotr: Ah, nice! ---- 2018-03-21 19:46:41 UTC - Piotr: Are Producer instances thread safe? ---- 2018-03-21 19:46:51 UTC - Piotr: (in the java client) ---- 2018-03-21 19:46:58 UTC - Matteo Merli: yes, all parts of API are thread safe ---- 2018-03-21 19:47:15 UTC - Piotr: super ---- 2018-03-21 20:32:30 UTC - Piotr: Is it a good idea to cache Producers? I will have very sporadical writes to my topics and creating a new producer with all the connection logic etc done every time seems like a waste - unless there is some caching done in the client already, or something already built I should use instead (currently doing it using guava CacheBuilder) ---- 2018-03-21 20:34:38 UTC - Matteo Merli: producer instances themselves are not cached, though the client instance is what holds the connection and thread pools ---- 2018-03-21 20:35:49 UTC - Piotr: alright so caching them does make sense then ---- 2018-03-21 20:35:55 UTC - Piotr: (in my case) ---- 2018-03-21 20:36:49 UTC - Matteo Merli: yes, also that once you have a producer instance it will be valid throughput the lifespan of your process ---- 2018-03-21 20:37:21 UTC - Piotr: is it safe to keep the PulsarClient around for the lifetime as well? Will it reconnect etc? ---- 2018-03-21 20:37:21 UTC - Matteo Merli: so you don’t need to invalidate that cache if there are errors ---- 2018-03-21 23:17:02 UTC - Igor Zubchenok: @Igor Zubchenok has joined the channel ---- 2018-03-22 00:31:26 UTC - Rabi Panda: Hi Everyone - while everyone is asking about all advanced stuff (atleast advanced from where I stand) - Can someone help with the setup? I haven't been able to get over the line with the steps on the Apache site ---- 2018-03-22 00:33:11 UTC - Rabi Panda: after untar - when I execute 'bin/pulsar standalone', I get the below - Caused by: java.io.IOException: Can not attach to current VM ---- 2018-03-22 00:35:11 UTC - Piotr: @Rabi Panda is that all? ---- 2018-03-22 00:36:38 UTC - Rabi Panda: 2018-03-22 00:28:02,898 - ERROR - [main:Journal$LastLogMark@226] - Problems reading from data/standalone/bookkeeper0/current/lastMark (this is okay if it is the first time starting this bookie 2018-03-22 00:28:05,523 - INFO - [bookie-reg-5-1-SendThread(127.0.0.1:2181):ClientCnxn$SendThread@1032] - Opening socket connection to server 127.0.0.1/127.0.0.1:2181. Will not attempt to authenticate using SASL (unknown error) 2018-03-22 00:28:05,524 - INFO - [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:59528 2018-03-22 00:28:05,524 - INFO - [bookie-reg-5-1-SendThread(127.0.0.1:2181):ClientCnxn$SendThread@876] - Socket connection established to 127.0.0.1/127.0.0.1:2181, initiating session 2018-03-22 00:28:05,531 - INFO - [ForceWriteThread:Journal$ForceWriteThread@457] - ForceWrite Thread started 2018-03-22 00:28:05,542 - INFO - [BookieJournal-3181:JournalChannel@143] - Opening journal data/standalone/bookkeeper0/current/1624b1b081a.txn 2018-03-22 00:28:05,667 - INFO - [bookie-reg-5-1-SendThread(127.0.0.1:2181):ClientCnxn$SendThread@1299] - Session establishment complete on server 127.0.0.1/127.0.0.1:2181, sessionid = 0x1624b1af9920002, negotiated timeout = 10000 WARNING: An illegal reflective access operation has occurred WARNING: Illegal reflective access by org.apache.bookkeeper.util.NativeIO (file:/opt/apache-pulsar-1.22.0-incubating/lib/org.apache.bookkeeper-bookkeeper-server-4.3.1.91-yahoo.jar) to field java.io.FileDescriptor.fd WARNING: Please consider reporting this to the maintainers of org.apache.bookkeeper.util.NativeIO WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations WARNING: All illegal access operations will be denied in a future release ---- 2018-03-22 00:37:01 UTC - Rabi Panda: @Piotr Hi Piotr - some more errors! ---- 2018-03-22 00:38:42 UTC - Igor Zubchenok: Hi there! Please suggest a best practice or antipatterns guides about using Pulsar? Thanks) heavy_plus_sign : Piotr ---- 2018-03-22 00:39:43 UTC - Piotr: @Rabi Panda I´m also quite new to pulsar, so just trying my best to help here knowing what I know. That IOException should have a stack trace somewhere. Did you just get that 1 line when trying to start pulsar, and is the other output from the log? ---- 2018-03-22 00:40:20 UTC - Piotr: (maybe you can find the rest of the IOException stack trace somewhere, which should explain the error). ---- 2018-03-22 00:41:55 UTC - Rabi Panda: @Piotr - All that I pasted were error messages that was thrown back when I kicked off the command
2018-03-22 00:28:05,524 - INFO - [NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@192] - Accepted socket connection from /127.0.0.1:59528 2018-03-22 00:28:05,524 - INFO - [bookie-reg-5-1-SendThread(127.0.0.1:2181):ClientCnxn$SendThread@876] - Socket connection established to 127.0.0.1/127.0.0.1:2181, initiating session 2018-03-22 00:28:05,531 - INFO - [ForceWriteThread:Journal$ForceWriteThread@457] - ForceWrite Thread started 2018-03-22 00:28:05,542 - INFO - [BookieJournal-3181:JournalChannel@143] - Opening journal data/standalone/bookkeeper0/current/1624b1b081a.txn 2018-03-22 00:28:05,667 - INFO - [bookie-reg-5-1-SendThread(127.0.0.1:2181):ClientCnxn$SendThread@1299] - Session establishment complete on server 127.0.0.1/127.0.0.1:2181, sessionid = 0x1624b1af9920002, negotiated timeout = 10000 WARNING: An illegal reflective access operation has occurred WARNING: Illegal reflective access by org.apache.bookkeeper.util.NativeIO (file:/opt/apache-pulsar-1.22.0-incubating/lib/org.apache.bookkeeper-bookkeeper-server-4.3.1.91-yahoo.jar) to field java.io.FileDescriptor.fd WARNING: Please consider reporting this to the maintainers of org.apache.bookkeeper.util.NativeIO WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations WARNING: All illegal access operations will be denied in a future release ---- 2018-03-22 00:41:56 UTC - Piotr: One possible reason for the VM connection to fail would be if the JVMs are ran by different users ---- 2018-03-22 00:43:38 UTC - Rabi Panda: am pretty sure I did not use a different user for anything! Is there a place where I can check? ---- 2018-03-22 00:45:30 UTC - Piotr: @Rabi Panda I´m afraid my helpfulness ends there :confused: You could check broker.log (maybe called pulsar.log) for more error messages ---- 2018-03-22 00:47:54 UTC - Rabi Panda: I was afraid you would say that @Piotr :slightly_smiling_face: I can't see any *.log files inside the pulsar installation. $ find . -name *.log ./data/standalone/bookkeeper0/current/ledgers/000003.log ./data/standalone/bookkeeper0/current/0.log ./data/standalone/bookkeeper0/current/locations/000003.log ---- 2018-03-22 00:48:40 UTC - Piotr: you should check /var/log , /usr/local/var/log etc ---- 2018-03-22 00:53:38 UTC - Rabi Panda: no luck there either - how did you setup? did you install it from a different location than this one? <https://pulsar.incubator.apache.org/download/> ---- 2018-03-22 00:54:44 UTC - Piotr: I installed it using homebrew, a package manager for mac os ---- 2018-03-22 00:55:11 UTC - Piotr: if you´re on mac, that´s what I would recommend for a dev setup ---- 2018-03-22 00:55:44 UTC - Rabi Panda: am on Ubuntu ---- 2018-03-22 00:57:45 UTC - Piotr: You could try installing the <http://streaml.io|streaml.io> sandbox (<https://streaml.io/docs/getting-started>) ---- 2018-03-22 00:57:54 UTC - Piotr: as pulsar is part of their stack ---- 2018-03-22 01:04:23 UTC - jia zhai: @Rabi Panda, try to see log4j settings if it is console, since it is started by systemd, you can probably get the log by journalctl, like: $ journalctl -u pulsar.service ---- 2018-03-22 01:06:55 UTC - Rabi Panda: Hi @jia zhai $ journalctl -u pulsar.service -- Logs begin at Wed 2018-03-21 23:34:59 GMT, end at Thu 2018-03-22 01:05:17 GMT. -- -- No entries -- ---- 2018-03-22 01:07:40 UTC - jia zhai: maybe `pulsar.service` is not the name in your running ---- 2018-03-22 01:12:12 UTC - jia zhai: try without `-u pulsar.service` to get the unit name of pulsar first? ---- 2018-03-22 01:14:00 UTC - Rabi Panda: only thing with pulsar on it - org.gnome.Shell.desktop[1671]: Window manager warning: 0x2200010 (Pulsar con) appears to be one of the offending windows with a timestamp of 1734526. Working around... ---- 2018-03-22 01:32:37 UTC - jia zhai: not sure how to do with gnome shell. maybe we could configure default log4j settings to a target it to log file? ---- 2018-03-22 01:36:04 UTC - jia zhai: There are docker files for Pulsar, maybe using docker would avoid some platform issues. ---- 2018-03-22 07:17:46 UTC - Sijie Guo: @Igor Zubchenok > Please suggest a best practice or antipatterns guides about using Pulsar? any more specific areas? ---- 2018-03-22 07:26:48 UTC - Sijie Guo: @Rabi Panda (/cc @Piotr @jia zhai) by default it should `bin/pulsar standalone` is outputting logs to console. are you running pulsar with java9? from the output messages, it seems like java 9. if that’s the case, it explains. because pulsar standalone is loading a aspectj-weaver agent for instrumenting zookeeper. I think jdk9 is deprecating dynamically self-attaching java agents. so that is probably the problem you are seeing. trying to install java 8 and try it again? +1 : jia zhai, Piotr ---- 2018-03-22 07:29:54 UTC - Sijie Guo: I filed an issue for that - <https://github.com/apache/incubator-pulsar/issues/1419> ---- 2018-03-22 07:46:07 UTC - Piotr: Hmm. I am getting "Error recovering ledger" and "Not enough non-faulty bookies available". There seems to be a bug issue about this started q few months ago with no action. Any ideas what this could be? I just left the project and dev server on overnight... ---- 2018-03-22 07:48:31 UTC - Sijie Guo: @Piotr - “Not enough non-faulty bookies available” means there is no enough bookies (storage node) available in the cluster to compromise the replication settings (by default I think it is 2). so the question is - how do you deploy pulsar? is it standalone? ---- 2018-03-22 07:48:37 UTC - Sijie Guo: or a cluster ---- 2018-03-22 07:51:57 UTC - Piotr: hmm that is odd. It is standalone on my dev machine with 580 GB available ---- 2018-03-22 07:54:45 UTC - Sijie Guo: hmm if it is standalone, then that’s even odd. can you find any errors in the log? or how did you run “standalone” overnight? ---- 2018-03-22 07:59:17 UTC - Piotr: There are some interesting things in the log: <https://pastebin.com/0vv4JgZt> ---- 2018-03-22 07:59:51 UTC - Piotr: Quite a few "Disconnected from bookie channel" and then "ERROR - [BookKeeperClientWorker-38-1:ReadLastConfirmedOp@127] - While readLastConfirmed ledger: 5 did not hear success responses from all quorums" ---- 2018-03-22 08:00:10 UTC - Piotr: I installed it using the homebrew formulae from stremlio ---- 2018-03-22 08:11:15 UTC - Sijie Guo: @Piotr interesting.. is your dev machine a laptop? did you close the laptop? or ? ---- 2018-03-22 08:16:52 UTC - Piotr: it is yes - hmm not sure, I might have. Could that have caused it to become corrupt you mean? ---- 2018-03-22 08:17:59 UTC - Piotr: it´s a macbook and I think mac os shuts down/pauses processes gracefully ---- 2018-03-22 08:18:53 UTC - Piotr: I´m sure I can fix this by reinstalling - but would be nice to know the reason for this happening - it´s a bit worrying to see this early on :wink: ---- 2018-03-22 08:19:32 UTC - Piotr: or can I do something about it using the admin? ---- 2018-03-22 08:24:51 UTC - Sijie Guo: it is a bit strange to me. standalone starts bookie and pulsar broker in the same jvm process. so if pulsar is alive, both broker and bookie should be alive. I am worried about os putting some threads into sleep state, which can cause some weird thread state. can you use “brew” to restart the service? - brew services restart pulsar ---- 2018-03-22 08:30:20 UTC - Piotr: I restarted it before and that didn´t help, but this time it did.. Hmm. Here is the full log since the restart: <https://pastebin.com/6yJyadMa> ---- 2018-03-22 08:31:03 UTC - Piotr: It´s working now, so no biggie. But feel free to check it out :slightly_smiling_face: And thanks for the help ---- 2018-03-22 08:32:13 UTC - Sijie Guo: @Piotr I will check the standalone code to see what can be improved +1 : Piotr ----