toasters November 1999

toasters@lists.teaparty.net

104 participants
95 discussions

Re: Java Thread & TimeDaemon
by Sam Cramer 02 Nov '99

02 Nov '99

> I've been running 5.3.2R1 since Aug 12th, and I had set up the timed to : > > timed.enable on > timed.log off > timed.max_skew 2s > timed.proto ntp > timed.sched hourly > timed.servers 205.246.15.5 > > Today, I started seeing in my syslogs : > > ..... Tue Nov 2 10:10:42 CST [Java Thread]: TimeDaemon: NTP server > '205.246.15.5' reports that the time is Tue Nov 02 10:14:45 CST 1999, > > for both of the toasters. Anyone know what this is about? these toasters > have been up for around 40 days or so... The syslog messages result from a misunderstanding regarding the meaning of the "timed.max_skew" option. Looking at the name of the option, one could reasonably assume that it controls how far out of whack we'll let the filer's notion of time get before we synchronize with the time server. In fact, it controls how far out of whack a timer server's time can be before the filer will no longer synchronize with the timer server. We use this field to prevent "time poisoning" of the filer by a grossly inaccurate time server. If we had it to do over again, we'd probably change the name of this option to something less confusing. So, when you specified "options timed.max_skew 2s", you were actually saying "if the time reported by the time server is more than 2 seconds away from the filer's notion of the correct time, don't synchronize to the time server." You can eliminate these messages, and get the filers back into synch with the time server by setting "timed.max_skew" to something like 30 minutes (30m). Sam

1 0

Java Thread & TimeDaemon
by Jay Orr 02 Nov '99

02 Nov '99

Huh.... I've been running 5.3.2R1 since Aug 12th, and I had set up the timed to : timed.enable on timed.log off timed.max_skew 2s timed.proto ntp timed.sched hourly timed.servers 205.246.15.5 Today, I started seeing in my syslogs : ..... Tue Nov 2 10:10:42 CST [Java Thread]: TimeDaemon: NTP server '205.246.15.5' reports that the time is Tue Nov 02 10:14:45 CST 1999, for both of the toasters. Anyone know what this is about? these toasters have been up for around 40 days or so... ----------- Jay Orr Systems Administrator Fujitsu Nexion Inc. St. Louis, MO

1 0

Re: General VIF questions
by Sam Cramer 01 Nov '99

01 Nov '99

Menachem> Hi, We don't "yet" have clustered systems, so maybe my question Menachem> is a bit off, but how do you read a partners /et/rc file in the Menachem> event of a crash - are they mirrored to each other? If so how Menachem> does this mechanism work? The disks are dual-ported, so each node has access to the partner's disks. Check out http://www.netapp.com/products/filer/clusters.html and http://www.netapp.com/tech_library/1004.html for more details on Netapp's clustering technology. Sam

1 0

"quota report" for users
by Brian Tao 01 Nov '99

01 Nov '99

Is there any plans to have "quota report" only look up and display a single user? I know you can do it for quota trees, but not for user or group quotas (according to the 5.2 and 5.3 man pages). I have this in my /etc/quotas file: /www1 tree 20g 1m /www2 tree 20g 1m * user@www1 1g * user@www2 1g I know about using the "quota" command from a UNIX NFS client, but that has some drawbacks (only seeing filesystems it has mounted, output formatting problems on Solaris, not being able to specify a particular mount point, etc.). I'd rather ask the filer how much disk space and inodes are owned by a giving uid, without having to go through the entire "quota report" listing (currently at 5500 lines and growing). -- Brian Tao (BT300, taob(a)risc.org) "Though this be madness, yet there is method in't"

1 0

Re: General VIF questions
by Sam Cramer 01 Nov '99

01 Nov '99

I wrote: > When a node takes over for its partner, it reads the partner's /etc/rc > file. The ifconfig commands are processed specially: they are examined > for a "partner" clause. If a partner clause is found, the interface > named in the partner clause takes over for the interface specified in the > ifconfig command. If no partner clause is found, the ifconfig command is > ignored. Whoops --- not enough caffeine! As tkaczma points out, this is exactly backwards: the partner clause is processed by the node destined to take over. Partner mappings (records which define which partner interface the local interface will take over for) --- are recorded at that time. When the node takes over its partner, the partner /etc/rc is run; partner ifconfig statements are matched with the partner mappings in order to determine which local interface will takeover for the partner interface or address in question. Sam

3 2

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters November 1999