AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:
CFG:
PT:0
S:andy.sharp@lsi.com
RQ:
SSV:mhbs.lsil.com
NSV:
SSH:
R:<Larry.Scheer@lsi.com>,<Chris.Vandever@lsi.com>,<Dave.Limato@lsi.com>,<John.Keiffer@lsi.com>,<Jan.Seidel@lsi.com>,<dl-qa@lsi.com>,<dl-engineering@lsi.com>
MAID:2
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/LSI/INBOX	0	DEC609CD0E54B2448DAF023C89AE9755E250CEF0@cosmail02.lsi.com
X-Sylpheed-End-Special-Headers: 1
Date: Tue, 29 Sep 2009 13:43:48 -0700
From: Andrew Sharp <andy.sharp@lsi.com>
To: "Scheer, Larry" <Larry.Scheer@lsi.com>
Cc: "Vandever, Chris" <Chris.Vandever@lsi.com>, "Limato, Dave"
 <Dave.Limato@lsi.com>, "Keiffer, John" <John.Keiffer@lsi.com>, "Seidel,
 Jan" <Jan.Seidel@lsi.com>, DL-ONStor-QA <dl-qa@lsi.com>,
 DL-ONStor-Engineering <dl-engineering@lsi.com>
Subject: Re: old elog files gone?
Message-ID: <20090929134348.5a30444a@ripper.onstor.net>
In-Reply-To: <DEC609CD0E54B2448DAF023C89AE9755E250CEF0@cosmail02.lsi.com>
References: <85A1D09038E3C1438820EF7A7FAFDD30F5102283@cosmail01.lsi.com>
	<D7A889C980962746B30DE07864593C02C7C986D6@cosmail02.lsi.com>
	<F9DCB1C30AC37B4EB352D0DE0AE18E5F9772CA43@cosmail02.lsi.com>
	<DEC609CD0E54B2448DAF023C89AE9755E250CEF0@cosmail02.lsi.com>
Organization: LSI
X-Mailer: Sylpheed-Claws 2.6.0 (GTK+ 2.8.20; x86_64-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

This command, and the specific way that it works, was requested by
either QA, Support or Ops, I don't remember which.  It's only legitimate
use, in my opinion, is the one that it was requested for.  If memory
serves, the idea that all prior history be wiped out was the
intention.  Perhaps the answer is that the tests not use this command
except for the small percentage of cases when that is the
intention as well.

On Tue, 29 Sep 2009 12:51:56 -0600 "Scheer, Larry"
<Larry.Scheer@lsi.com> wrote:

> Just a thought, don't know how valid it is... I seem to remember a
> somewhat common problem we had before the CF was reformatted to a new
> partition layout. The problem  was  having enough space on the CF for
> large cluster databases and /var filling up. The elog clear log
> command was a quick way to recover from this problem. Since the CF
> has been reformatted and we are using 1 GB flashes on cougars I don't
> here of this problem so much.
>=20
> Larry
> ________________________________________
> From: Vandever, Chris
> Sent: Tuesday, September 29, 2009 11:27 AM
> To: Limato, Dave; Keiffer, John; Seidel, Jan; DL-ONStor-QA;
> DL-ONStor-Engineering Subject: RE: old elog files gone?
>=20
> I know of no good reason to delete ANY log entries.  They provide
> history of what happened on the system and are one of our best
> diagnostic tools.  Deleting any log entries make it harder to
> diagnose problems.
>=20
> Perhaps the real answer is to remove the "elog clear log" command
> entirely.
>=20
> ChrisV
> ________________________________
> From: Limato, Dave
> Sent: Tuesday, September 29, 2009 11:25 AM
> To: Keiffer, John; Seidel, Jan; Vandever, Chris; DL-ONStor-QA;
> DL-ONStor-Engineering Subject: RE: old elog files gone?
>=20
> I think this was purposefully implemented this way. I don't know why
> but I think it was.
>=20
> I mean clear log, means clear the logs right? Maybe to your point we
> need a roll logs option.
>=20
> ________________________________
> From: Keiffer, John
> Sent: Tuesday, September 29, 2009 11:20 AM
> To: Seidel, Jan; Vandever, Chris; DL-ONStor-QA; DL-ONStor-Engineering
> Subject: RE: old elog files gone?
>=20
> I found the offending script:
>=20
> [root@c13r34-rhel4 test]# grep "elog clear
> log" ../Build/test-logs/t/all/features/mirror/local/workflow/* ../Build/t=
est-logs/t/all/features/mirror/local/workflow/047_LocalMirror_Workflow.t.10=
.2.62.9.log:17:08:32
> g9r62 diag> elog clear log
>=20
> Shall I file a defect so we can track it?
>=20
> Also, doesn=E2=80=99t anyone think that =E2=80=98elog clear log=E2=80=99 =
could be made a
> little less dangerous? Maybe it should only empty the messages file
> without touching all the others. I can=E2=80=99t think why we would want =
to
> blow all the older messages away. If we really needed that, we could
> add a =E2=80=98-all=E2=80=99 option or something to the command? What say=
 ye?
> ECR/Defect?
>=20
> Thanks,
> John
> 408-376-3106
> ________________________________
> From: Seidel, Jan
> Sent: Tuesday, September 29, 2009 11:13 AM
> To: Keiffer, John; Vandever, Chris; DL-ONStor-QA;
> DL-ONStor-Engineering Subject: RE: old elog files gone?
>=20
> The Test Harness itself is not clearing any elogs, but there are some
> specific test cases which do this. I=E2=80=99m working on removing this, =
but
> it=E2=80=99s not that trivial (the test cases obviously clear the elogs f=
or a
> reason, so I need to replace this code with something else).
>=20
> An =E2=80=9Celog clear log=E2=80=9D removes ALL elogs (including the mess=
ages.0 and
> so on). So that=E2=80=99s probably the reason why you have these files
> missing.
>=20
> Regards,
> Jan
>=20
> From: Keiffer, John
> Sent: Tuesday, September 29, 2009 11:09 AM
> To: Vandever, Chris; DL-ONStor-QA; DL-ONStor-Engineering
> Subject: RE: old elog files gone?
>=20
> Good point. I suppose clearing an elog is okay, but removing old ones
> would be bad.
>=20
> I=E2=80=99m not sure what happened though. I=E2=80=99m going to look into=
 the TH logs
> and see if I find anything interesting=E2=80=A6
>=20
> Thanks,
> John
> 408-376-3106
> ________________________________
> From: Vandever, Chris
> Sent: Tuesday, September 29, 2009 11:06 AM
> To: Keiffer, John; DL-ONStor-QA; DL-ONStor-Engineering
> Subject: RE: old elog files gone?
>=20
> It used to be if you ran anything from the automated test harness
> that it cleared the elogs before it started.  Don't know if that's
> still true or not.  Obviously, this is suboptimal.
>=20
> ChrisV
> ________________________________
> From: Keiffer, John
> Sent: Tuesday, September 29, 2009 10:44 AM
> To: DL-ONStor-QA; DL-ONStor-Engineering
> Subject: old elog files gone?
> I=E2=80=99ve been running R4.1.0.0DEV-CGDBG-092109 for a while. Last nigh=
t a
> had an FP crash (and probably more after that).
>=20
> I just went to copy the elog files to the defect dir, and I found
> that I only had a /var/log/onstor/messages file and no .0 or gz
> backup files.
>=20
> g9r62:~# l /var/log/onstor/
> total 2.1M
> -rw-r--r-- 1 root root    0 Sep 19  2008 vscan
> -rw-r--r-- 1 root root 525K Aug  7 13:30 clusServ.log.0
> drwxr-xr-x 4 root root 1.0K Aug 19 14:29 samba/
> -rw-r--r-- 1 root root  16K Sep 10 11:57 sys_upgrade.log
> -rw-r--r-- 1 root root   66 Sep 28 06:25 elogs_mod_time_stamp
> -rw-r--r-- 1 root root  174 Sep 29 09:14 clusServ.log.old
> drwxr-xr-x 7 root root 1.0K Sep 29 09:17 ../
> drwxr-xr-x 3 root root 1.0K Sep 29 09:17 ./
> -rw-r--r-- 1 root root 1.5M Sep 29 10:04 messages
> -rw-r--r-- 1 root root  334 Sep 29 10:30 clusServ.log
>=20
> Has this happened to anyone else? What caused this?
>=20
> Thank you,
> John Keiffer
> 408-376-3106
> LSI - QA Engineer
> john.keiffer@lsi.com<mailto:john.keiffer@lsi.com>
>=20
