Mailing-List: contact java-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: java-user@lucene.apache.org
Received-SPF: pass (athena.apache.org: domain of
 prakash.chinnakannan@gmail.com designates 209.85.215.49 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <CAL8PwkbjVSTMg3GfFn3xXn-kPShXyspwYT201eJURYa0kRnWXA@mail.gmail.com>
References: 
 <CACDAp1CyqYmuj507RVMAEv2VNfKTTLgGXvV4pEZKdwK_3yaYow@mail.gmail.com>
	<CAL8PwkYLvxKy4SsdjXp6QEPjkP7c+0JozT-kq=+mVjQhjnC0kg@mail.gmail.com>
	<CACDAp1AVg040Zu+-W97jzzu5ngXRpB7xLog-5yUkbyNB6jvojA@mail.gmail.com>
	<CAL8PwkbjVSTMg3GfFn3xXn-kPShXyspwYT201eJURYa0kRnWXA@mail.gmail.com>
Date: Mon, 29 Jul 2013 14:33:31 +0530
Message-ID: 
 <CACDAp1BAeF2=vrb+A9PmkUFKy85OhjOezTZJ0_SP+MXhHFZ=wg@mail.gmail.com>
Subject: Re: ERROR: could not read any segments file in directory
From: Prakash Chinnakannan <prakash.chinnakannan@gmail.com>
To: java-user@lucene.apache.org
Content-Type: multipart/alternative; boundary=001a11c296543f5c1d04e2a2c4a3

--001a11c296543f5c1d04e2a2c4a3
Content-Type: text/plain; charset=ISO-8859-1

Thanks Mike, we'd restored the index content from the backup.  I will also
suggest to move the index to local storage.

> Do you commit frequently and has this index been running for a very, very
long time?

Yes, index files would be there for more than 3 yrs and frequent add/update
and delete operations are going on over that.

>What is that INDEX_REQUEST file doing in the index?

Not sure how it got created.. will check that Mike.

Thanks,
Prakash


On Sat, Jul 27, 2013 at 4:09 PM, Michael McCandless <
lucene@mikemccandless.com> wrote:

> OK, you have only one segments file, and it sounds like it was
> corrupted by the crash of your SAN.  I don't think there's much you
> can do but re-index.
>
> Maybe move your index to local storage: it sounds like this SAN is not
> to be trusted.
>
> Separately, your segment numbers are truly immense.  Do you commit
> frequently and has this index been running for a very, very long time?
>
> What is that INDEX_REQUEST file doing in the index?
>
> Mike McCandless
>
> http://blog.mikemccandless.com
>
>
> On Fri, Jul 26, 2013 at 1:57 PM, Prakash Chinnakannan
> <prakash.chinnakannan@gmail.com> wrote:
> > Thanks for your time Mike,
> >
> > Yes the commit has been made successfully before crashing.
> >
> > Here is the output of ls -lrt on searchIndex/ directory
> >
> >>>
> >
> > searchIndex#] ls -lrt
> > total 50624356
> > -rw-r--r-- 1 root root 32991 Jul 26 06:42 _18vyk7.fnm
> > -rw-r--r-- 1 root root 39608652 Jul 26 06:42 _18vyk7.fdx
> > -rw-r--r-- 1 root root 2585873905 Jul 26 06:42 _18vyk7.fdt
> > -rw-r--r-- 1 root root 3746137075 Jul 26 07:04 _18vyk7.tis
> > -rw-r--r-- 1 root root 48247452 Jul 26 07:04 _18vyk7.tii
> > -rw-r--r-- 1 root root 31312798534 Jul 26 07:04 _18vyk7.prx
> > -rw-r--r-- 1 root root 7066658159 Jul 26 07:04 _18vyk7.frq
> > -rw-r--r-- 1 root root 6976073133 Jul 26 07:05 _18vyk7.nrm
> > -rw-r--r-- 1 root root 1551 Jul 26 07:13 _18vyk7_1f.del
> > -rw-r--r-- 1 root root 138 Jul 26 07:13 INDEX_REQUEST
> > -rw-r--r-- 1 root root 10134260 Jul 26 07:13 _18wlug.cfs
> > -rw-r--r-- 1 root root 20 Jul 26 07:13 segments.gen
> > -rw-r--r-- 1 root root 5141 Jul 26 07:13 segments_174k7h
> > -rw-r--r-- 1 root root 563 Jul 26 07:13 _18wmdw.tis
> > -rw-r--r-- 1 root root 35 Jul 26 07:13 _18wmdw.tii
> > -rw-r--r-- 1 root root 32 Jul 26 07:13 _18wmdw.prx
> > -rw-r--r-- 1 root root 32 Jul 26 07:13 _18wmdw.frq
> > -rw-r--r-- 1 root root 12 Jul 26 07:13 _18wmdw.fdx
> > -rw-r--r-- 1 root root 266 Jul 26 07:13 _18wmdw.fdt
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdv.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdt.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmds.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdr.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdq.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdp.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdo.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdn.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdm.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdl.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdk.cfs
> > -rw-r--r-- 1 root root 1149834 Jul 26 07:13 _18wmdj.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdi.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdh.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdg.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdf.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmde.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdd.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdc.cfs
> > -rw-r--r-- 1 root root 35454 Jul 26 07:13 _18wmdb.cfs
> > -rw-r--r-- 1 root root 51334 Jul 26 07:13 _18wmda.cfs
> > -rw-r--r-- 1 root root 1117999 Jul 26 07:13 _18wmcz.cfs
> > <<
> >
> > Below is the another exception trace, which happens when we try to
> > initialize the IndexWriter in our application.
> >
> >
> >>>
> > ERROR   13748   [SearchSe] (): [2013-07-26 12:14:10,911] The Lucene
> > index at location [searchIndex] is corrupt. Please wipe away the
> > entire directory or replace with a back up. It's not reasonable to
> > continue.
> > org.apache.lucene.index.CorruptIndexException: Unknown format version:
> > -318130790
> >         at
> org.apache.lucene.index.SegmentInfos.read(SegmentInfos.java:272)
> >         at
> org.apache.lucene.index.IndexFileDeleter.<init>(IndexFileDeleter.java:168)
> >         at
> org.apache.lucene.index.IndexWriter.<init>(IndexWriter.java:1172)
> > <<
> >
> > Thanks,
> > Prakash
> >
> >
> >
> > On Fri, Jul 26, 2013 at 8:50 PM, Michael McCandless <
> > lucene@mikemccandless.com> wrote:
> >
> >> Likely there's nothing easy you can do to recover the index.
> >>
> >> If the crash was merely an "outage", and the IO system did not flip
> >> bits on files that were committed, then the index should have been
> >> intact.
> >>
> >> Can you post the ls -l of the index directory?
> >>
> >> Had you successfully committed to this index before crashing?
> >>
> >> Mike McCandless
> >>
> >> http://blog.mikemccandless.com
> >>
> >>
> >> On Fri, Jul 26, 2013 at 10:59 AM, Prakash Chinnakannan
> >> <prakash.chinnakannan@gmail.com> wrote:
> >> > Hi,
> >> >
> >> > Today we'd the SAN outage and it looks the lucene index directory got
> >> > corrupted. We tried to fix it by using CheckIndex and below is the
> >> > exception trace.
> >> >
> >> > Do we've any other possible ways to recover the index contents?
> >> >
> >> > ~#] java -cp lucene-3.2.0.jar org.
> >> > apache.lucene.index.CheckIndex searchIndex/ -fix
> >> >
> >> > NOTE: testing will be more thorough if you run java with
> >> > '-ea:org.apache.lucene...', so assertions are enabled
> >> >
> >> > Opening index @ searchIndex/
> >> >
> >> > ERROR: could not read any segments file in directory
> >> > org.apache.lucene.index.CorruptIndexException: Unknown format version:
> >> > -318130790
> >> > at org.apache.lucene.index.SegmentInfos.read(SegmentInfos.java:272)
> >> > at
> org.apache.lucene.index.SegmentInfos$1.doBody(SegmentInfos.java:358)
> >> > at
> >>
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:753)
> >> > at
> >>
> org.apache.lucene.index.SegmentInfos$FindSegmentsFile.run(SegmentInfos.java:592)
> >> > at org.apache.lucene.index.SegmentInfos.read(SegmentInfos.java:354)
> >> > at org.apache.lucene.index.CheckIndex.checkIndex(CheckIndex.java:319)
> >> > at org.apache.lucene.index.CheckIndex.main(CheckIndex.java:940)
> >> >
> >> > Thanks for your help.
> >> >
> >> > --
> >> > Prakash C
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> >> For additional commands, e-mail: java-user-help@lucene.apache.org
> >>
> >>
> >
> >
> > --
> > Prakash C
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>


-- 
Prakash C

--001a11c296543f5c1d04e2a2c4a3--