This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2016-03-23
Channels
- # admin-announcements (6)
- # aleph (3)
- # beginners (38)
- # boot (119)
- # braid-chat (15)
- # braveandtrue (1)
- # clara (4)
- # cljs-dev (56)
- # cljsfiddle (12)
- # cljsjs (15)
- # cljsrn (6)
- # clojars (4)
- # clojure (113)
- # clojure-art (1)
- # clojure-berlin (1)
- # clojure-dusseldorf (3)
- # clojure-india (15)
- # clojure-new-zealand (3)
- # clojure-poland (1)
- # clojure-russia (83)
- # clojure-uk (18)
- # clojurescript (97)
- # community-development (9)
- # cursive (1)
- # data-science (1)
- # datomic (12)
- # emacs (14)
- # hoplon (350)
- # immutant (2)
- # jobs (2)
- # jobs-discuss (23)
- # keechma (74)
- # liberator (1)
- # off-topic (1)
- # om (127)
- # onyx (54)
- # parinfer (74)
- # pedestal (1)
- # proton (5)
- # re-frame (6)
- # reagent (4)
- # remote-jobs (17)
- # ring-swagger (1)
- # slack-help (5)
- # untangled (16)
- # yada (21)
Hi guys, are there a way to ask datomic about address of secondary (passive) instance of HA transactors? I've got a way to get Active instance details:
(datomic.peer/transactor-endpoint db-uri)
I can't find a doc for datomic.peer lib for take a look on all methods
The active transactor writes its location into storage for Peers to find it. I don't think the passive transactor(s) record their location until they become active.
stuartsierra: so I haven't another way except "get active and compare"?
I do not know of any mechanism in Datomic to find passive transactors.
feature request
@lowl4tency: the issue here is that the passive transactor doesn’t write to storage, so there’s no way for a peer to get to it.
If this is re: cloud formation ops - i.e., which instance to kill, you should kill at the cloud formation level. I.e., put up a new transactor pair, then when you see HeartMonitorMsec
sum go up, kill the old transactor pair.
This will force a failover to the new transactor pair (they’re ready to take over when they’re reporting HeartMonitorMsec
metrics. Depending on whether or not there’s a standby at present (again discernible if anything is posting that metric), you can either wait for HeartMonitorMsec
to show up again after being absent, or look at samples/count or sum values which will tell you that more than 1 standby transactor is ready to go.
Thank you
I am trying to alter my schema. In order to prepare the addition of a unique constraint I added an index with something along the lines of
[{:db/id :person/external-id
:db/index true
:db.alter/_attribute :db.part/db}]
But now my transactor is dying with:
WARN default datomic.update - {:message "Index creation failed", :db-id “store-0b6b5518-0141-4392-b077-1729ea4464c7", :pid 1965, :tid 12}
java.io.IOException: No such file or directory
at java.io.UnixFileSystem.createFileExclusively(Native Method) ~[na:1.8.0_72]
at java.io.File.createTempFile(File.java:2024) ~[na:1.8.0_72]
at datomic.external_sort$temp_file_io$reify__2757.make_temp_file(external_sort.clj:22) ~[datomic-transactor-pro-0.9.5350.jar:na]
at datomic.external_sort$file_system_sorter$fn__2850.invoke(external_sort.clj:113) ~[datomic-transactor-pro-0.9.5350.jar:na]
I am not sure what I am doing wrong here. I assume permissions for the default temp directory are wrong even though it does not seem like they are and /tmp
exists. What is confusing to me is why this has not been an issue before. Does Datomic only create these temp files when I alter an existing schema to add an index?