9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
* [9fans] Reliability of fossil
@ 2003-10-22 13:13 arisawa
  2003-10-22 15:09 ` Fco.J.Ballesteros
  0 siblings, 1 reply; 3+ messages in thread
From: arisawa @ 2003-10-22 13:13 UTC (permalink / raw)
  To: 9fans

Hello,

My comments about reliability of fossil.

I have two fossil server. One is for home network, another is for
Internet.

CASE fossil for home network
That works perfectly fine. No trouble since the first installation
term% ls /n/dump/*
d-r-xr-xr-x M 29 adm adm 0 Jan 18  2003 /n/dump/2003/0118
d-r-xr-xr-x M 29 adm adm 0 Jan 19  2003 /n/dump/2003/0119
d-r-xr-xr-x M 29 adm adm 0 Jan 19  2003 /n/dump/2003/0119.1
d-r-xr-xr-x M 29 adm adm 0 Jan 20  2003 /n/dump/2003/0120
d-r-xr-xr-x M 29 adm adm 0 Feb  4  2003 /n/dump/2003/0204
d-r-xr-xr-x M 29 adm adm 0 Feb  6  2003 /n/dump/2003/0206
...
d-r-xr-xr-x M 29 adm adm 0 Oct 19 22:25 /n/dump/2003/1019
d-r-xr-xr-x M 29 adm adm 0 Oct 20 07:46 /n/dump/2003/1020
d-r-xr-xr-x M 29 adm adm 0 Oct 20 22:18 /n/dump/2003/1020.1
d-r-xr-xr-x M 29 adm adm 0 Oct 21 22:35 /n/dump/2003/1021
d-r-xr-xr-x M 29 adm adm 0 Oct 22 09:03 /n/dump/2003/1022
term%

CASE fossil for Internet
My computers are placed at my office of my university.
Three of my machines were dead at 8 August because of typhoon.
Windows were open and the machines were drenched!
File server and CPU server were alive.
So I reconstructed file server to be auth+venti+fossil server.
The system continued working without any trouble until 30 September.
My system, ar.aichi-u.ac.jp (with alias plan9.aichi-u.ac.jp) has been
in service for educational purpose. A massive amount of accesses
comes from students, specially at every Tuesday.
The day, 30 September, was the first Tuesday of this semester.
 From that day, my fossil falls into broken state at every Tuesday.
I suspected hardwares because they were old and I replaced them
with new one. However even after the replacement, fossil continues
getting
into trouble at every Tuesday.
Now I have added new machine that is based on kfs for the services to
students.

I appreciate the effort of Russ. Fossil is really fine. But it is
desired to
test under hard condition.

Kenji Arisawa



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [9fans] Reliability of fossil
  2003-10-22 13:13 [9fans] Reliability of fossil arisawa
@ 2003-10-22 15:09 ` Fco.J.Ballesteros
  2003-10-23  8:00   ` arisawa
  0 siblings, 1 reply; 3+ messages in thread
From: Fco.J.Ballesteros @ 2003-10-22 15:09 UTC (permalink / raw)
  To: 9fans

[-- Attachment #1: Type: text/plain, Size: 141 bytes --]

how did you reconstruct the file server? copying the arenas back?
Did you start with a brand new fossil?
Is this fossil the one breaking?

[-- Attachment #2: Type: message/rfc822, Size: 3703 bytes --]

From: arisawa@ar.aichi-u.ac.jp
To: 9fans@cse.psu.edu
Subject: [9fans] Reliability of fossil
Date: Wed, 22 Oct 2003 22:13:38 +0900
Message-ID: <8CFD66E7-0491-11D8-A23F-000393A941BC@ar.aichi-u.ac.jp>

Hello,

My comments about reliability of fossil.

I have two fossil server. One is for home network, another is for
Internet.

CASE fossil for home network
That works perfectly fine. No trouble since the first installation
term% ls /n/dump/*
d-r-xr-xr-x M 29 adm adm 0 Jan 18  2003 /n/dump/2003/0118
d-r-xr-xr-x M 29 adm adm 0 Jan 19  2003 /n/dump/2003/0119
d-r-xr-xr-x M 29 adm adm 0 Jan 19  2003 /n/dump/2003/0119.1
d-r-xr-xr-x M 29 adm adm 0 Jan 20  2003 /n/dump/2003/0120
d-r-xr-xr-x M 29 adm adm 0 Feb  4  2003 /n/dump/2003/0204
d-r-xr-xr-x M 29 adm adm 0 Feb  6  2003 /n/dump/2003/0206
...
d-r-xr-xr-x M 29 adm adm 0 Oct 19 22:25 /n/dump/2003/1019
d-r-xr-xr-x M 29 adm adm 0 Oct 20 07:46 /n/dump/2003/1020
d-r-xr-xr-x M 29 adm adm 0 Oct 20 22:18 /n/dump/2003/1020.1
d-r-xr-xr-x M 29 adm adm 0 Oct 21 22:35 /n/dump/2003/1021
d-r-xr-xr-x M 29 adm adm 0 Oct 22 09:03 /n/dump/2003/1022
term%

CASE fossil for Internet
My computers are placed at my office of my university.
Three of my machines were dead at 8 August because of typhoon.
Windows were open and the machines were drenched!
File server and CPU server were alive.
So I reconstructed file server to be auth+venti+fossil server.
The system continued working without any trouble until 30 September.
My system, ar.aichi-u.ac.jp (with alias plan9.aichi-u.ac.jp) has been
in service for educational purpose. A massive amount of accesses
comes from students, specially at every Tuesday.
The day, 30 September, was the first Tuesday of this semester.
 From that day, my fossil falls into broken state at every Tuesday.
I suspected hardwares because they were old and I replaced them
with new one. However even after the replacement, fossil continues
getting
into trouble at every Tuesday.
Now I have added new machine that is based on kfs for the services to
students.

I appreciate the effort of Russ. Fossil is really fine. But it is
desired to
test under hard condition.

Kenji Arisawa

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [9fans] Reliability of fossil
  2003-10-22 15:09 ` Fco.J.Ballesteros
@ 2003-10-23  8:00   ` arisawa
  0 siblings, 0 replies; 3+ messages in thread
From: arisawa @ 2003-10-23  8:00 UTC (permalink / raw)
  To: 9fans

Hello,

> how did you reconstruct the file server? copying the arenas back?
> Did you start with a brand new fossil?
> Is this fossil the one breaking?
>
I reconstructed venti and fossil from the scratch using the copy in kfs:
- first time (at the time my system is damaged by typhoon)
	I constructed auth+venti+fossil server from old machine that had been
my file server.
	HDs were old 8GB + 8GB.
- second time ( at the first corruption of fossil )
	I reconstructed auth+venti+fossil server using the copy in kfs, not
from arenas.
	Although I suspected the origin of the trouble is in old machine:
		P3/Xeon/866 with 256MB memory
- third time (at second corruption of fossil )
	I replaced the old machine with new machine (Athron 2200+ with 512MB
memory) but 	HDs were still old.
	I reconstructed the server using the copy in kfs, not from arenas.
	Although I suspected the origin of the trouble is in old HDs.
- fourth time (at third trouble of fossil )
	I replaced old HDs with new HD (80GB).
	arenas and fossil are constructed from the scratch.
	I expected fossil can run without trouble.
- fifth time (at fourth trouble of fossil that were happened at mid
night dump)
	I simply execute:
	checkarenas
	checkindex
	flchk
- sixth time (at fifth trouble of fossil )
	I simply execute:
	checkarenas
	checkindex
	flchk
	and posted the problems to 9fans

The current archives are:
d-r-xr-xr-x M 31 adm adm 0 Oct 18 16:00 /n/temp/2003/1018
d-r-xr-xr-x M 31 adm adm 0 Oct 19 04:00 /n/temp/2003/1019
d-r-xr-xr-x M 31 adm adm 0 Oct 20 09:06 /n/temp/2003/1020
d-r-xr-xr-x M 31 adm adm 0 Oct 20 09:49 /n/temp/2003/1020.1
d-r-xr-xr-x M 31 adm adm 0 Oct 21 04:00 /n/temp/2003/1021
d-r-xr-xr-x M 31 adm adm 0 Oct 22 04:00 /n/temp/2003/1022
d-r-xr-xr-x M 31 adm adm 0 Oct 23 04:00 /n/temp/2003/1023

In constructing arenas and fossil I got new version of fossil and venti
binary
from sources.cs.bell-labs.com and used them as it is.

They are:
--rwxrwxr-x M 4568 sys    sys   1497 Sep  7 10:00 /386/bin/fossil/conf
--rwxrwxr-x M 4568 sys    sys 232263 Sep 27 02:42 /386/bin/fossil/flchk
--rwxrwxr-x M 4568 sys    sys 230228 Sep 27 02:42 /386/bin/fossil/flfmt
--rwxrwxr-x M 4568 sys    sys 345764 Oct 14 11:22 /386/bin/fossil/fossil
--rwxrwxr-x M 4568 sys sys 160540 Sep 27 02:46 /386/bin/venti/buildindex
--rwxrwxr-x M 4568 sys sys 164405 Sep 27 02:46
/386/bin/venti/checkarenas
--rwxrwxr-x M 4568 sys sys 161346 Sep 27 02:46 /386/bin/venti/checkindex
--rwxrwxr-x M 4568 sys sys 150452 Sep 27 02:46 /386/bin/venti/clumpstats
--rwxrwxr-x M 4568 sys sys   1506 Jun 23 19:33 /386/bin/venti/conf
--rwxrwxr-x M 4568 sys sys 104558 Sep 27 02:46 /386/bin/venti/copy
--rwxrwxr-x M 4568 sys sys 131327 Sep 27 02:46 /386/bin/venti/findscore
--rwxrwxr-x M 4568 sys sys 132572 Sep 27 02:46 /386/bin/venti/fmtarenas
--rwxrwxr-x M 4568 sys sys 150815 Sep 27 02:46 /386/bin/venti/fmtindex
--rwxrwxr-x M 4568 sys sys 150699 Sep 27 02:46 /386/bin/venti/fmtisect
--rwxrwxr-x M 4568 sys sys 131051 Sep 27 02:46 /386/bin/venti/rdarena
--rwxrwxr-x M 4568 sys sys 100831 Sep 27 02:46 /386/bin/venti/read
--rwxrwxr-x M 4568 sys sys  99981 Sep 27 02:46 /386/bin/venti/sync
--rwxrwxr-x M 4568 sys sys 175875 Sep 27 02:46 /386/bin/venti/syncindex
--rwxrwxr-x M 4568 sys sys 251676 Sep 27 02:46 /386/bin/venti/venti
--rwxrwxr-x M 4568 sys sys 104568 Sep 27 02:46
/386/bin/venti/verifyarena
--rwxrwxr-x M 4568 sys sys 181861 Sep 27 02:46 /386/bin/venti/wrarena
--rwxrwxr-x M 4568 sys sys 100783 Sep 27 02:46 /386/bin/venti/write

Thanks nemo for your interest.

Kenji Arisawa



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2003-10-23  8:00 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-10-22 13:13 [9fans] Reliability of fossil arisawa
2003-10-22 15:09 ` Fco.J.Ballesteros
2003-10-23  8:00   ` arisawa

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).