I am getting an error on my workstations saying: The Semaphore timeout period has expired! The workstations can not access the server(unix sco server). Its is not always the same workstations that have this problem and it has been happening every week. Any suggestions or solutions?
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations gkittelson on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Semaphore Timeout expired? (SCO UNIX Server)
- Thread starter Grassland
- Start date
stanhubble
MIS
Do you have enough semaphores defined for the application?
what is the application?
what is the version and flavour of sco? (Openserver vs Unixware)
Most of the doc's and suggestions for semaphore values are from a time when 32 MB of ram was expensive.
what is the application?
what is the version and flavour of sco? (Openserver vs Unixware)
Most of the doc's and suggestions for semaphore values are from a time when 32 MB of ram was expensive.
- Thread starter
- #3
stanhubble
MIS
sar -m is a good starting point.
ipcs -a will tell you about current use of semaphores and message queues.
/etc/conf/cf.d/stune will contain the kernel parameters that affect semaphores (SEM*) if they have been changed from the default.
I am not familiar with esker's VSIFAX and their online doc's are next to useless but,
for our database application we modify the following:
The memory hit for the increased structures is negligible on a current system.
ipcs -a will tell you about current use of semaphores and message queues.
/etc/conf/cf.d/stune will contain the kernel parameters that affect semaphores (SEM*) if they have been changed from the default.
I am not familiar with esker's VSIFAX and their online doc's are next to useless but,
for our database application we modify the following:
Code:
current default max
SEMMNI 400 10 8192
SEMMNU 1200 30 8192
SEMMSL 200 25 150 (it complains about
this one, but it
still works)
SEMMNS 1200 60 8192The memory hit for the increased structures is negligible on a current system.
- Thread starter
- #5
You can also use the sysdef command to see the current kernel parameters. To only view SEMAPHORE values,
# sysdef|grep SEM
We use VSI*FAX on several of our Clients' systems, but I haven't seen the error you mention. Most of our VSI*FAX installations are all UNIX (not using their Client/Server connection method). Have you looked through the VSI*FAX log files on the server (/usr/vsifax3/spool/logs) to see if there are any additional clues?
# sysdef|grep SEM
We use VSI*FAX on several of our Clients' systems, but I haven't seen the error you mention. Most of our VSI*FAX installations are all UNIX (not using their Client/Server connection method). Have you looked through the VSI*FAX log files on the server (/usr/vsifax3/spool/logs) to see if there are any additional clues?
stanhubble
MIS
How many users are you trying to serve?
the ml350 g2 is just fine for most applications. you need to look at what is happening on the system. VSIFAX and ftp may just be showing the symptoms or results of something else going wrong or set inappropriately.
step 1: find out what is actually happening
what is the output of sar with no options?
what is the output of sar -r ?
does a console session seem sluggish?
do you have all the patches loaded?
do you have the compaq efs loaded? what version?
etc....etc....
there are lots of reasons for a slow/non-responsive system. the key is finding out where the bottleneck is? CPU, memory, or i/o
the ml350 g2 is just fine for most applications. you need to look at what is happening on the system. VSIFAX and ftp may just be showing the symptoms or results of something else going wrong or set inappropriately.
step 1: find out what is actually happening
what is the output of sar with no options?
what is the output of sar -r ?
does a console session seem sluggish?
do you have all the patches loaded?
do you have the compaq efs loaded? what version?
etc....etc....
there are lots of reasons for a slow/non-responsive system. the key is finding out where the bottleneck is? CPU, memory, or i/o
Stan presents a good point:
If your results of the "sar" command show the system is very busy, you can run this command to see which process is taking up more than it's share of CPU resources:
ps -e -o "pcpu" -o "pid=" -o "time" -o "user=" -o "args=" |
awk '$1 > 5' |
sort -r -n
Run it a few different times to see if a particular process continues to appear. This command basically shows any process which consumes more than 5% of the CPU at any specific moment.
If your results of the "sar" command show the system is very busy, you can run this command to see which process is taking up more than it's share of CPU resources:
ps -e -o "pcpu" -o "pid=" -o "time" -o "user=" -o "args=" |
awk '$1 > 5' |
sort -r -n
Run it a few different times to see if a particular process continues to appear. This command basically shows any process which consumes more than 5% of the CPU at any specific moment.
- Status
Similar threads
- Locked
- Question