Hi Henrik
Here's an interesting bug. Or at least I think it's a bug. I set the MAXMSG_CLIENT to 1024, and it fixed the problem I was having. (See below)
Later on, I had to restart the server, and it crapped out. I remember that sometimes deleting ~hobbit/server/tmp/* before a startup can fix things, so I gave it a try. Still no go. I then restarted it with --debug, and the only log file getting written to was hobbitlaunch.log (See extract below)
I then reversed the only thing that had changed - removed the MAXMSG_CLIENT entry, and it all started up fine.
A bug? Or a PEBKAC?
Regards Vernon
--- snip --- 2006-08-07 16:23:54 Loading hostnames 2006-08-07 16:23:54 Opening file /export/home/hobbit//server/etc/bb-hosts 2006-08-07 16:23:54 Opening file /export/home/hobbit//server/etc/client-local.cfg 2006-08-07 16:23:54 Loading saved state 2006-08-07 16:23:54 Cannot access checkpoint file /export/home/hobbit//server/tmp/hobbitd.chk for restore 2006-08-07 16:23:54 Setting up network listener on 0.0.0.0:1984 2006-08-07 16:23:54 Setting up signal handlers 2006-08-07 16:23:54 Setting up hobbitd channels 2006-08-07 16:23:54 Setting up status channel (id=1) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',1) 2006-08-07 16:23:54 ftok() returns: 0x100748F 2006-08-07 16:23:54 shmget() returns: 0x1200001E 2006-08-07 16:23:54 Setting up stachg channel (id=2) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',2) 2006-08-07 16:23:54 ftok() returns: 0x200748F 2006-08-07 16:23:54 shmget() returns: 0x1200001F 2006-08-07 16:23:54 Setting up page channel (id=3) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',3) 2006-08-07 16:23:54 ftok() returns: 0x300748F 2006-08-07 16:23:54 shmget() returns: 0x12000020 2006-08-07 16:23:54 Setting up data channel (id=4) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',4) 2006-08-07 16:23:54 ftok() returns: 0x400748F 2006-08-07 16:23:54 shmget() returns: 0x12000021 2006-08-07 16:23:54 Setting up notes channel (id=5) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',5) 2006-08-07 16:23:54 ftok() returns: 0x500748F 2006-08-07 16:23:54 shmget() returns: 0x12000022 2006-08-07 16:23:54 Setting up enadis channel (id=6) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',6) 2006-08-07 16:23:54 ftok() returns: 0x600748F 2006-08-07 16:23:54 shmget() returns: 0x12000023 2006-08-07 16:23:54 Setting up client channel (id=7) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',7) 2006-08-07 16:23:54 ftok() returns: 0x700748F 2006-08-07 16:23:54 shmget() returns: 0x12000024 2006-08-07 16:23:54 Setting up clichg channel (id=8) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',8) 2006-08-07 16:23:54 ftok() returns: 0x800748F 2006-08-07 16:23:54 Could not get shm of size 1048576: Invalid argument 2006-08-07 16:23:54 Cannot setup clichg channel --- snip ---
-----Original Message----- From: Everett, Vernon [mailto:Vernon.Everett at dpi.wa.gov.au] Sent: Monday, 7 August 2006 12:17 PM To: hobbit at hswn.dk Subject: RE: [hobbit] Message to long
Yep, that was it. Thanks Lars
Cheers Vernon
-----Original Message----- From: lars ebeling [mailto:lars.ebeling at leopg9.no-ip.org] Sent: Monday, 7 August 2006 11:53 AM To: hobbit at hswn.dk Subject: Re: [hobbit] Message to long
Have a look at hobbitserver.cfg manpage. MAXMSG_CLIENT
Regards Lars
----- Original Message ----- From: "Everett, Vernon" <Vernon.Everett at dpi.wa.gov.au> To: <hobbit at hswn.dk> Sent: Monday, August 07, 2006 3:42 AM Subject: [hobbit] Message to long
Hi all
Despite my best efforts, my procs page remains purple. I found this in the hobbitd page - which also remains yellow. (Server name changed to pluto)
--- snip --- Latest errormessages: Oversize data/client msg from 10.1.5.242 truncated (n=949733, limit 524288) First line: pluto|sunos|sunos Oversize stachg msg from 10.1.5.242 for pluto:ports truncated (n=508447, limit=262144) Oversize page/ack/notify msg from 10.1.5.242 for pluto:ports truncated (n=508456, limit=262144) Oversize status msg from 10.1.5.242 for pluto:ports truncated (n=508452, limit=262144) Oversize data/client msg from 10.1.5.242 truncated (n=977049, limit 524288) First line: pluto|sunos|sunos Oversize status msg from 10.1.5.242 for pluto:ports truncated (n=508448, limit=262144) --- snip ---
I am assuming that the output from the ports and procs test is too large. How do we make Hobbit happy with the larger message? This machine is a proxy server, so we can expect many procs and a long ports output.
Regards Vernon
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
MAXMSG_CLIENT seems to need to allocate a shared memory segment of equal size. If your kernel shared memory settings are set too low (eg. if you are running defaults on solaris) then you may not be able to allocate the segment. If that happens, things like you just described seem to happen...
you could possibly have dangling segments left over that were not cleaned up when you restarted hobbit, or possibly another app.
Try checking what is allocated with "ipcs -a", and increasing the parameter "shmmax".
This is a PRIVATE message. If you are not the intended recipient, please delete without copying and kindly advise us by e-mail of the mistake in delivery. NOTE: Regardless of content, this e-mail shall not operate to bind CSC to any order or other contract unless pursuant to explicit written agreement or government initiative expressly permitting the use of e-mail for such purpose.
"Everett, Vernon" <Vernon.Everett at dpi.wa.gov.au> 07/08/2006 09:40 Please respond to hobbit at hswn.dk
To <hobbit at hswn.dk> cc
Subject RE: [hobbit] Message to long
Hi Henrik
Here's an interesting bug. Or at least I think it's a bug. I set the MAXMSG_CLIENT to 1024, and it fixed the problem I was having. (See below)
Later on, I had to restart the server, and it crapped out. I remember that sometimes deleting ~hobbit/server/tmp/* before a startup can fix things, so I gave it a try. Still no go. I then restarted it with --debug, and the only log file getting written to was hobbitlaunch.log (See extract below)
I then reversed the only thing that had changed - removed the MAXMSG_CLIENT entry, and it all started up fine.
A bug? Or a PEBKAC?
Regards Vernon
--- snip --- 2006-08-07 16:23:54 Loading hostnames 2006-08-07 16:23:54 Opening file /export/home/hobbit//server/etc/bb-hosts 2006-08-07 16:23:54 Opening file /export/home/hobbit//server/etc/client-local.cfg 2006-08-07 16:23:54 Loading saved state 2006-08-07 16:23:54 Cannot access checkpoint file /export/home/hobbit//server/tmp/hobbitd.chk for restore 2006-08-07 16:23:54 Setting up network listener on 0.0.0.0:1984 2006-08-07 16:23:54 Setting up signal handlers 2006-08-07 16:23:54 Setting up hobbitd channels 2006-08-07 16:23:54 Setting up status channel (id=1) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',1) 2006-08-07 16:23:54 ftok() returns: 0x100748F 2006-08-07 16:23:54 shmget() returns: 0x1200001E 2006-08-07 16:23:54 Setting up stachg channel (id=2) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',2) 2006-08-07 16:23:54 ftok() returns: 0x200748F 2006-08-07 16:23:54 shmget() returns: 0x1200001F 2006-08-07 16:23:54 Setting up page channel (id=3) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',3) 2006-08-07 16:23:54 ftok() returns: 0x300748F 2006-08-07 16:23:54 shmget() returns: 0x12000020 2006-08-07 16:23:54 Setting up data channel (id=4) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',4) 2006-08-07 16:23:54 ftok() returns: 0x400748F 2006-08-07 16:23:54 shmget() returns: 0x12000021 2006-08-07 16:23:54 Setting up notes channel (id=5) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',5) 2006-08-07 16:23:54 ftok() returns: 0x500748F 2006-08-07 16:23:54 shmget() returns: 0x12000022 2006-08-07 16:23:54 Setting up enadis channel (id=6) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',6) 2006-08-07 16:23:54 ftok() returns: 0x600748F 2006-08-07 16:23:54 shmget() returns: 0x12000023 2006-08-07 16:23:54 Setting up client channel (id=7) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',7) 2006-08-07 16:23:54 ftok() returns: 0x700748F 2006-08-07 16:23:54 shmget() returns: 0x12000024 2006-08-07 16:23:54 Setting up clichg channel (id=8) 2006-08-07 16:23:54 calling ftok('/export/home/hobbit//server',8) 2006-08-07 16:23:54 ftok() returns: 0x800748F 2006-08-07 16:23:54 Could not get shm of size 1048576: Invalid argument 2006-08-07 16:23:54 Cannot setup clichg channel --- snip ---
-----Original Message----- From: Everett, Vernon [mailto:Vernon.Everett at dpi.wa.gov.au] Sent: Monday, 7 August 2006 12:17 PM To: hobbit at hswn.dk Subject: RE: [hobbit] Message to long
Yep, that was it. Thanks Lars
Cheers Vernon
-----Original Message----- From: lars ebeling [mailto:lars.ebeling at leopg9.no-ip.org] Sent: Monday, 7 August 2006 11:53 AM To: hobbit at hswn.dk Subject: Re: [hobbit] Message to long
Have a look at hobbitserver.cfg manpage. MAXMSG_CLIENT
Regards Lars
----- Original Message ----- From: "Everett, Vernon" <Vernon.Everett at dpi.wa.gov.au> To: <hobbit at hswn.dk> Sent: Monday, August 07, 2006 3:42 AM Subject: [hobbit] Message to long
Hi all
Despite my best efforts, my procs page remains purple. I found this in the hobbitd page - which also remains yellow. (Server name changed to pluto)
--- snip --- Latest errormessages: Oversize data/client msg from 10.1.5.242 truncated (n=949733, limit 524288) First line: pluto|sunos|sunos Oversize stachg msg from 10.1.5.242 for pluto:ports truncated (n=508447, limit=262144) Oversize page/ack/notify msg from 10.1.5.242 for pluto:ports truncated (n=508456, limit=262144) Oversize status msg from 10.1.5.242 for pluto:ports truncated (n=508452, limit=262144) Oversize data/client msg from 10.1.5.242 truncated (n=977049, limit 524288) First line: pluto|sunos|sunos Oversize status msg from 10.1.5.242 for pluto:ports truncated (n=508448, limit=262144) --- snip ---
I am assuming that the output from the ports and procs test is too large. How do we make Hobbit happy with the larger message? This machine is a proxy server, so we can expect many procs and a long ports output.
Regards Vernon
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to hobbit-unsubscribe at hswn.dk
participants (2)
-
cspargo2@csc.com
-
Vernon.Everett@dpi.wa.gov.au