All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Shu, Xinxin" <xinxin.shu@intel.com>
To: Robert LeBlanc <robert@leblancnet.us>,
	ceph-devel <ceph-devel@vger.kernel.org>
Subject: RE: Need help compiling 0.94.1 with jemalloc
Date: Tue, 19 May 2015 07:38:31 +0000	[thread overview]
Message-ID: <75674D092A819E4189E91166C74CB90D01553437@shsmsx102.ccr.corp.intel.com> (raw)
In-Reply-To: <CAANLjFq1xNGms4q811oJp7TtzLjqWRQO9t-gcNdkvr5vWHP5GQ@mail.gmail.com>

Please refer to https://github.com/facebook/rocksdb/blob/master/include/rocksdb/options.h#L560 , the latest rocksdb move these options to BlockBasedTableOption, I think you should try latest ceph, this PR https://github.com/ceph/ceph/pull/4622 has been merged into master

Cheers,
xinxin

-----Original Message-----
From: ceph-devel-owner@vger.kernel.org [mailto:ceph-devel-owner@vger.kernel.org] On Behalf Of Robert LeBlanc
Sent: Tuesday, May 19, 2015 12:24 PM
To: ceph-devel
Subject: Need help compiling 0.94.1 with jemalloc

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

I'm trying to build Ceph 0.94.1 with jemalloc and without tcmalloc.
I've downloaded rocksdb from their master branch (to get the jemalloc
option) and built rocksdb with the shared libraries. When I try and build Ceph, it fails on:

os/RocksDBStore.cc: In member function 'int RocksDBStore::do_open(std::ostream&, bool)':
os/RocksDBStore.cc:65:15: error: 'struct rocksdb::Options' has no member named 'block_cache'
     ldoptions.block_cache = rocksdb::NewLRUCache(options.cache_size);
               ^
os/RocksDBStore.cc:68:15: error: 'struct rocksdb::Options' has no member named 'block_size'
     ldoptions.block_size = options.block_size;
               ^
os/RocksDBStore.cc:72:15: error: 'struct rocksdb::Options' has no member named 'filter_policy'
     ldoptions.filter_policy = _filterpolicy;
               ^
os/RocksDBStore.cc:86:15: error: 'struct rocksdb::Options' has no member named 'block_restart_interval'
     ldoptions.block_restart_interval = options.block_restart_interval;
               ^

I can see in the config.log that it is finding the shared rockdb libraries ok. I've configured with:

./configure $(for i in /usr/{lib64,lib}/jvm/java/include{,/linux}; do [ -d $i ] && java_inc="$java_inc -I$i"; done) --prefix=/usr --localstatedir=/var --sysconfdir=/etc --docdir=/usr/share/doc/ceph --with-nss --without-cryptopp --with-rest-bench --with-debug --enable-cephfs-java --with-librocksdb-static=no --with-radosgw --with-jemalloc --without-tcmalloc --with-librocksdb

There is one warning in the config.log, but it seems to happily move right along.

configure:21346: checking for LIBROCKSDB
configure:21353: $PKG_CONFIG --exists --print-errors "rocksdb"
Package rocksdb was not found in the pkg-config search path.
Perhaps you should add the directory containing `rocksdb.pc'
to the PKG_CONFIG_PATH environment variable No package 'rocksdb' found
configure:21356: $? = 1
configure:21370: $PKG_CONFIG --exists --print-errors "rocksdb"
Package rocksdb was not found in the pkg-config search path.
Perhaps you should add the directory containing `rocksdb.pc'
to the PKG_CONFIG_PATH environment variable No package 'rocksdb' found
configure:21373: $? = 1
configure:21387: result: no
No package 'rocksdb' found
configure:21417: checking for open in -lrocksdb
configure:21442: gcc -o conftest -g -O2   conftest.c -lrocksdb   >&5
configure:21442: $? = 0
configure:21451: result: yes
configure:21678: checking for io_submit in -laio
configure:21703: gcc -o conftest -g -O2   conftest.c -laio  -lrocksdb  >&5
configure:21703: $? = 0
configure:21712: result: yes

I don't have a rocksdb.pc from my build of it, so I'm not sure what exactly it is talking about.

I've had some very interesting results from preloading different memory allocators and I'd like to try a complete jemalloc install then share my results. So any push in the right direction would be much appreciated.

Thank you,

- ----------------
Robert LeBlanc
GPG Fingerprint 79A2 9CA4 6CC4 45DD A904  C70E E654 3BB2 FA62 B9F1 -----BEGIN PGP SIGNATURE-----
Version: Mailvelope v0.13.1
Comment: https://www.mailvelope.com
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=EXv4
-----END PGP SIGNATURE-----
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@vger.kernel.org More majordomo info at  http://vger.kernel.org/majordomo-info.html

  reply	other threads:[~2015-05-19  7:38 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-19  4:23 Need help compiling 0.94.1 with jemalloc Robert LeBlanc
2015-05-19  7:38 ` Shu, Xinxin [this message]
2015-05-20  0:56   ` Robert LeBlanc

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=75674D092A819E4189E91166C74CB90D01553437@shsmsx102.ccr.corp.intel.com \
    --to=xinxin.shu@intel.com \
    --cc=ceph-devel@vger.kernel.org \
    --cc=robert@leblancnet.us \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.