| When you issue the kill command and process number, you are also sending a signal associated with the kill. We did not specify a signal in our kill example; however, the default signal of 15, or SIGTERM, was used. These signals are used by the system to communicate with processes. The signal of 15 we used to terminate our process is a software termination signal that is usually enough to terminate a user process such as the find we had started. A process that is difficult to kill may require the SIGKILL, or 9 signal. This signal causes an immediate termination of the process. I use this only as a last resort because processes killed with SIGKILL do not always terminate smoothly. To kill such processes as the shell, you sometimes have to use SIGKILL. You can use either the signal name or number. These signal numbers sometimes vary from system to system, so view the manual page for signal, usually in section 5, to see the list of signals on your system. A list of some of the most frequently used signal numbers and corresponding signals follows: | Signal Number | Signal |
|---|
| 1 | SIGHUP | | 2 | SIGINT | | 3 | SIGQUIT | | 9 | SIGKILL | | 15 | SIGTERM | | 24 | SIGSTOP |
To kill a process with id 234 with SIGKILL, you would issue the following command: $ kill -9 234 | | | | | |> process id (PID) | |> signal number |> kill command to terminate the process Showing Remote Mounts with showmount showmount is used to show all remote systems (clients) that have mounted a local file system. showmount is useful for determining the file systems that are most often mounted by clients with NFS. The output of showmount is particularly easy to read because it lists the host name and directory that was mounted by the client. NFS servers often end up serving many NFS clients that were not originally intended to be served. This situation ends up consuming additional UNIX system resources on the NFS server, as well as additional network bandwidth. Keep in mind that any data transferred from an NFS server to an NFS client consumes network bandwidth, and in some cases, may be a substantial amount of bandwith if large files or applications are being transferred from the NFS server to the client. The following example is a partial output of showmount taken from a system. showmount runs on the HP-UX, AIX, and Linux systems I have been using throughout this chapter, but not on the Solaris system: # showmount -a sys100.ct.mp.com:/applic sys101.ct.mp.com:/applic sys102.cal.mp.com:/applic sys103.cal.mp.com:/applic sys104.cal.mp.com:/applic sys105.cal.mp.com:/applic sys106.cal.mp.com:/applic sys107.cal.mp.com:/applic sys108.cal.mp.com:/applic sys109.cal.mp.com:/applic sys200.cal.mp.com:/usr/users sys201.cal.mp.com:/usr/users sys202.cal.mp.com:/usr/users # showmount -a sys203.cal.mp.com:/usr/users sys204.cal.mp.com:/usr/users sys205.cal.mp.com:/usr/users sys206.cal.mp.com:/usr/users sys207.cal.mp.com:/usr/users sys208.cal.mp.com:/usr/users sys209.cal.mp.com:/usr/users The three following options are available for the showmount command: | -a | prints output in the format "name:directory," as shown above. | | -d | lists all the local directories that have been remotely mounted by clients. | | -e | prints a list of exported file systems. |
The following are examples of showmount -d and showmount -e: # showmount -d /applic /usr/users /usr/oracle /usr/users/emp.data /network/database /network/users /tmp/working # showmount -e export list for server101.cal.mp.com # showmount -e /applic /usr/users /cdrom Showing System Swap If your system has insufficient main memory for all the information it needs to work with, it will move pages of information to your swap area or swap entire processes to your swap area. Pages that were most recently used are kept in main memory, and those not recently used will be the first moved out of main memory. System administrators spend a lot of time determining the right amount of swap space for their systems. Insufficient swap may prevent a system from starting additional processes, hang applications, or not permit additional users to get access to the system. Having sufficient swap prevents these problems from occurring. System administrators usually go about determining the right amount of swap by considering many important factors, including the following: How much swap is recommended by the application(s) you run? Use the swap size recommended by your applications. Application vendors tend to be realistic when recommending swap space. There is sometimes competition among application vendors to claim the lowest memory and CPU requirements in order to keep the overall cost of solutions as low as possible, but swap space recommendations are usually realistic. How many applications will you run simultaneously? If you are running several applications, sum the swap space recommended for each application you plan to run simultaneously. If you have a database application that recommends 200 MBytes of swap and a development tool that recommends 100 MBytes of swap, then configure your system with 300 MBytes of swap, minimum. Will you be using substantial system resources on peripheral functionality such as NFS? The nature of NFS is to provide access to file systems, some of which may be very large, so this use may have an impact on your swap space requirements.
Swap is listed and manipulated on different UNIX variants with different commands. The following example shows listing the swap area on a Solaris system with swap -l: # swap -l swapfile dev swaplo blocks free /dev/dsk/c0t3d0s1 32,25 8 263080 209504 These values are all in 512 KByte blocks. In this case, the free blocks are 209504, which is a significant amount of the overall swap allocated on the system. You can view the amount of swap being consumed on your HP-UX system with swapinfo. The following is an example output of swapinfo: # swapinfo Kb Kb Kb PCT START/ Kb TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME dev 49152 10532 38620 21% 0 - 1 /dev/vg00/lvol2 dev 868352 10888 759160 1% 0 - 1 /dev/vg00/lvol8 reserve - 532360 -532360 memory 816360 469784 346576 58% Following is a brief overview of what swapinfo gives you. In the preceding example, the "TYPE" field indicated whether the swap was "dev" for device, "reserve" for paging space on reserve, or "memory." Memory is a way to allow programs to reserve more virtual memory than you have hard disk paging space set up for on your system. "Kb AVAIL" is the total swap space available in 1024-byte blocks. This includes both used and unused swap. "Kb USED" is the current number of 1024-byte blocks in use. "Kb FREE" is the difference between "Kb AVAIL" and "Kb USED." "PCT USED" is "Kb USED" divided by "Kb AVAIL." "START/LIMIT" is the block address of the start of the swap area. "Kb RESERVE" is "-" for device swap or the number of 1024-byte blocks for file system swap. "PRI" is the priority given to this swap area. "NAME" is the device name of the swap device. You can also issue the swapinfo command with a series of options. Here are some of the options you can include: | -m | to display output of swapinfo in MBytes rather than in 1024-byte blocks. | | -d | prints information related to device swap areas only. | | -f | prints information about file system swap areas only. |
sar: The System Activity Reporter sar is another UNIX command for gathering information about activities on your system. You can gather data over an extended time period with sar and later produce reports based on the data. sar is similar among UNIX variants in that the options and outputs are similar. The Linux system I was using for the examples did not support sar, but the Solaris, HP-UX, and AIX systems had the same options and nearly identical outputs. The following are some useful options to sar, along with examples of reports produced with these options where applicable: | sar -o | Save data in a file specified by "o." After the file name, you would usually also enter the time interval for samples and the number of samples. The following example shows saving the binary data in file /tmp/sar.data at an interval of 60 seconds 300 times: # sar -o /tmp/sar.data 60 300 The data in /tmp/sar.data can later be extracted from the file. | | sar -f | Specify a file from which you will extract data. | | sar -u | Report CPU utilization with the headings %usr, %sys, %wio, %idle with some processes waiting for block I/O, %idle. This report is similar to the iostat and vmstat CPU reports. You extract the binary data saved in a file to get CPU information, as shown in the following example. The following is a sar -u example: |
# sar -u -f /tmp/sar.data Header Information for your system 12:52:04 %usr %sys %wio %idle 12:53:04 62 4 5 29 12:54:04 88 5 3 4 12:55:04 94 5 1 0 12:56:04 67 4 4 25 12:57:04 59 4 4 32 12:58:04 61 4 3 32 12:59:04 65 4 3 28 13:00:04 62 5 16 17 13:01:04 59 5 9 27 13:02:04 71 4 3 22 13:03:04 60 4 4 32 13:04:04 71 5 4 20 13:05:04 80 6 8 7 13:06:04 56 3 3 37 13:07:04 57 4 4 36 13:08:04 66 4 4 26 13:09:04 80 10 2 8 13:10:04 73 10 2 15 13:11:04 64 6 3 28 13:12:04 56 4 3 38 13:13:04 55 3 3 38 13:14:04 57 4 3 36 13:15:04 70 4 5 21 13:16:04 65 5 9 21 13:17:04 62 6 2 30 13:18:04 60 5 3 33 13:19:04 77 3 4 16 13:20:04 76 5 3 15 . . . 14:30:04 50 6 6 38 14:31:04 57 12 19 12 14:32:04 51 8 20 21 14:33:04 41 4 9 46 14:34:04 43 4 9 45 14:35:04 38 4 6 53 14:36:04 38 9 7 46 14:37:04 46 3 11 40 14:38:04 43 4 7 46 14:39:04 37 4 5 54 14:40:04 33 4 5 58 14:41:04 40 3 3 53 14:42:04 44 3 3 50 14:43:04 27 3 7 64 Average 57 5 8 30 | sar -b | Report buffer cache activity. A database application such as Oracle would recommend that you use this option to see the effectiveness of buffer cache use. You extract the binary data saved in a file to get CPU information, as shown in the following example: |
# sar -b -f /tmp/sar.data Header information for your system 12:52:04 bread/s lread/s %rcache bwrit/s lwrit/s %wcache pread/s pwrit/s 12:53:04 5 608 99 1 11 95 0 0 12:54:04 7 759 99 0 14 99 0 0 12:55:04 2 1733 100 4 24 83 0 0 12:56:04 1 836 100 1 18 96 0 0 12:57:04 0 623 100 2 21 92 0 0 12:58:04 0 779 100 1 16 96 0 0 12:59:04 0 1125 100 0 14 98 0 0 13:00:04 2 1144 100 9 89 89 0 0 13:01:04 10 898 99 11 76 86 0 0 13:02:04 0 1156 100 0 14 99 0 0 13:03:04 1 578 100 2 22 88 0 0 13:04:04 5 1251 100 0 12 99 0 0 13:05:04 3 1250 100 0 12 97 0 0 13:06:04 1 588 100 0 12 98 0 0 13:07:04 1 649 100 2 15 86 0 0 13:08:04 1 704 100 2 15 86 0 0 13:09:04 1 1068 100 0 18 100 0 0 13:10:04 0 737 100 1 44 99 0 0 13:11:04 0 735 100 1 13 95 0 0 13:12:04 0 589 100 1 15 93 0 0 13:13:04 0 573 100 0 16 99 0 0 13:14:04 1 756 100 1 16 91 0 0 13:15:04 1 1092 100 9 49 81 0 0 13:16:04 2 808 100 6 82 93 0 0 13:17:04 0 712 100 1 9 93 0 0 13:18:04 1 609 100 0 13 97 0 0 13:19:04 1 603 100 0 10 99 0 0 13:20:04 0 1127 100 0 14 98 0 0 . . . 14:30:04 2 542 100 1 22 94 0 0 14:31:04 10 852 99 12 137 92 0 0 14:32:04 2 730 100 10 190 95 0 0 14:33:04 4 568 99 2 26 91 0 0 14:34:04 4 603 99 1 13 91 0 0 14:35:04 1 458 100 1 13 89 0 0 14:36:04 13 640 98 1 24 98 0 0 14:37:04 21 882 98 1 18 95 0 0 14:38:04 7 954 99 0 19 98 0 0 14:39:04 3 620 100 1 11 94 0 0 14:40:04 3 480 99 2 15 85 0 0 14:41:04 1 507 100 0 9 98 0 0 14:42:04 1 1010 100 1 10 91 0 0 14:43:04 5 547 99 1 9 93 0 0 Average 3 782 100 3 37 91 0 0 | sar -d | Report disk activity. You get the device name, percent that the device was busy, average number of requests outstanding for the device, number of data transfers per second for the device, and other information. You extract the binary data saved in a file to get CPU information, as shown in the following example: |
# sar -d -f /tmp/sar.data Header information for your system 12:52:04 device %busy avque r+w/s blks/s avwait avserv 12:53:04 c0t6d0 0.95 1.41 1 10 16.76 17.28 c5t4d0 100.00 1.03 20 320 8.36 18.90 c4t5d1 10.77 0.50 13 214 5.02 18.44 c5t4d2 0.38 0.50 0 3 4.61 18.81 12:54:04 c0t6d0 0.97 1.08 1 11 10.75 14.82 c5t4d0 100.00 1.28 54 862 9.31 20.06 c4t5d1 12.43 0.50 15 241 5.21 16.97 c5t4d2 0.37 0.50 0 3 3.91 18.20 12:55:04 c0t6d0 1.77 1.42 1 22 13.32 14.16 c5t4d0 100.00 0.79 26 421 8.33 16.00 c4t5d1 14.47 0.51 17 270 5.30 13.48 c5t4d2 0.72 0.50 0 7 4.82 15.69 12:56:04 c0t6d0 1.07 21.57 1 22 72.94 19.58 c5t4d0 100.00 0.60 16 251 6.80 13.45 c4t5d1 8.75 0.50 11 177 5.05 10.61 c5t4d2 0.62 0.50 0 6 4.79 15.43 12:57:04 c0t6d0 0.78 1.16 1 9 13.53 14.91 c5t4d0 100.00 0.66 15 237 7.60 13.69 c4t5d1 9.48 0.54 13 210 5.39 13.33 c5t4d2 0.87 0.50 1 10 4.86 14.09 12:58:04 c0t6d0 1.12 8.29 1 17 54.96 14.35 c5t4d0 100.00 0.60 11 176 7.91 14.65 c4t5d1 5.35 0.50 7 111 5.23 10.35 c5t4d2 0.92 0.50 1 10 4.63 16.08 12:59:04 c0t6d0 0.67 1.53 1 8 18.03 16.05 c5t4d0 99.98 0.54 11 174 7.69 14.09 c4t5d1 3.97 0.50 5 83 4.82 9.54 c5t4d2 1.05 0.50 1 11 4.69 16.29 13:00:04 c0t6d0 3.22 0.67 3 39 8.49 16.53 c5t4d0 100.00 0.60 65 1032 8.46 14.83 c4t5d1 21.62 0.50 31 504 5.30 8.94 c5t4d2 6.77 0.50 5 78 4.86 14.09 13:01:04 c0t6d0 4.45 3.08 5 59 25.83 11.49 c5t4d0 100.00 0.65 42 676 7.85 14.52 c4t5d1 21.34 0.55 30 476 5.87 18.49 c5t4d2 4.37 0.50 3 51 5.32 13.50 . . . 14:42:04 c0t6d0 0.53 0.83 0 7 12.21 16.33 c5t4d0 100.00 0.56 7 107 6.99 14.65 c4t5d1 6.38 0.50 7 113 4.97 15.18 c5t4d2 0.15 0.50 0 2 4.53 16.50 14:43:04 c0t6d0 0.52 0.92 0 7 11.50 15.86 c5t4d0 99.98 0.92 17 270 8.28 18.64 c4t5d1 10.26 0.50 9 150 5.35 16.41 c5t4d2 0.12 0.50 0 1 5.25 14.45 Average c0t6d0 1.43 108.80 2 26 0.00 14.71 Average c5t4d0 100.00 0.74 25 398 7.83 -10.31 Average c4t5d1 19.11 0.51 25 399 5.26 -13.75 Average c5t4d2 1.71 0.53 1 21 5.29 13.46 | sar -q | Report average queue length. You may have a problem any time the run queue length is greater than the number of processors on the system: |
# sar -q -f /tmp/sar.data Header information for your system 12:52:04 runq-sz %runocc swpq-sz %swpocc 12:53:04 1.1 20 0.0 0 12:54:04 1.4 51 0.0 0 12:55:04 1.3 71 0.0 0 12:56:04 1.1 22 0.0 0 12:57:04 1.3 16 0.0 0 12:58:04 1.1 14 0.0 0 12:59:04 1.2 12 0.0 0 13:00:04 1.2 21 0.0 0 13:01:04 1.1 18 0.0 0 13:02:04 1.3 20 0.0 0 13:03:04 1.2 15 0.0 0 13:04:04 1.2 20 0.0 0 13:05:04 1.2 43 0.0 0 13:06:04 1.1 14 0.0 0 13:07:04 1.2 15 0.0 0 13:08:04 1.2 26 0.0 0 13:09:04 1.5 38 0.0 0 13:10:04 1.5 30 0.0 0 13:11:04 1.2 23 0.0 0 13:12:04 1.3 11 0.0 0 13:13:04 1.3 12 0.0 0 13:14:04 1.4 16 0.0 0 13:15:04 1.4 27 0.0 0 13:16:04 1.5 20 0.0 0 13:17:04 1.3 21 0.0 0 13:18:04 1.1 15 0.0 0 13:19:04 1.2 19 0.0 0 13:20:04 1.4 22 0.0 0 . . . 14:30:04 1.5 5 0.0 0 14:31:04 1.6 12 0.0 0 14:32:04 1.4 9 0.0 0 14:33:04 1.1 6 0.0 0 14:34:04 1.3 3 0.0 0 14:35:04 1.1 4 0.0 0 14:36:04 1.2 6 0.0 0 14:37:04 1.4 5 0.0 0 14:38:04 1.2 10 0.0 0 14:39:04 1.3 4 0.0 0 14:40:04 1.1 3 0.0 0 14:41:04 1.6 3 0.0 0 14:42:04 1.1 4 0.0 0 14:43:04 1.3 1 0.0 0 Average 1.3 17 1.2 0 | sar -w | Report system swapping activity. |
# sar -w -f /tmp/sar.data Header information for your system 12:52:04 swpin/s bswin/s swpot/s bswot/s pswch/s 12:53:04 1.00 0.0 1.00 0.0 231 12:54:04 1.00 0.0 1.00 0.0 354 12:55:04 1.00 0.0 1.00 0.0 348 12:56:04 1.00 0.0 1.00 0.0 200 12:57:04 1.00 0.0 1.00 0.0 277 12:58:04 1.00 0.0 1.00 0.0 235 12:59:04 1.02 0.0 1.02 0.0 199 13:00:04 0.78 0.0 0.78 0.0 456 13:01:04 1.00 0.0 1.00 0.0 435 13:02:04 1.02 0.0 1.02 0.0 216 13:03:04 0.98 0.0 0.98 0.0 204 13:04:04 1.02 0.0 1.02 0.0 239 13:05:04 1.00 0.0 1.00 0.0 248 13:06:04 0.97 0.0 0.97 0.0 170 13:07:04 1.00 0.0 1.00 0.0 166 13:08:04 1.02 0.0 1.02 0.0 209 13:09:04 0.98 0.0 0.98 0.0 377 13:10:04 1.00 0.0 1.00 0.0 200 13:11:04 1.00 0.0 1.00 0.0 192 13:12:04 0.87 0.0 0.87 0.0 187 13:13:04 0.93 0.0 0.93 0.0 172 13:14:04 1.00 0.0 1.00 0.0 170 13:15:04 1.00 0.0 1.00 0.0 382 13:16:04 1.00 0.0 1.00 0.0 513 13:17:04 1.00 0.0 1.00 0.0 332 13:18:04 1.00 0.0 1.00 0.0 265 13:19:04 1.02 0.0 1.02 0.0 184 13:20:04 0.98 0.0 0.98 0.0 212 . . . 14:30:04 0.00 0.0 0.00 0.0 301 14:31:04 0.00 0.0 0.00 0.0 566 14:32:04 0.00 0.0 0.00 0.0 539 14:33:04 0.00 0.0 0.00 0.0 400 14:34:04 0.00 0.0 0.00 0.0 242 14:35:04 0.00 0.0 0.00 0.0 286 14:36:04 0.00 0.0 0.00 0.0 295 14:37:04 0.00 0.0 0.00 0.0 249 14:38:04 0.00 0.0 0.00 0.0 300 14:39:04 0.00 0.0 0.00 0.0 296 14:40:04 0.00 0.0 0.00 0.0 419 14:41:04 0.00 0.0 0.00 0.0 234 14:42:04 0.00 0.0 0.00 0.0 237 14:43:04 0.00 0.0 0.00 0.0 208 Average 0.70 0.0 0.70 0.0 346 Using timex to Analyze a Command If you have a specific command you want to find out more about, you can use timex, which reports the elapsed time, user time, and system time spent in the execution of any command you specify. timex is a good command for users because it gives you an idea of the system resources you are consuming when issuing a command. The following two examples show issuing timex with no options to get a short output of the amount of cpu consumed; the second example shows issuing timex -s to report "total" system activity on a Solaris system: martyp $ timex listing real 0.02 user 0.00 sys 0.02 martyp $ timex -s listing real 0.02 user 0.00 sys 0.01 SunOS 5.7 Generic sun4m 08/21 07:48:30 %usr %sys %wio %idle 07:48:31 32 68 0 0 07:48:30 bread/s lread/s %rcache bwrit/s lwrit/s %wcache pread/s pwrit/s 07:48:31 0 0 100 0 0 100 0 0 Average 0 0 100 0 0 100 0 0 07:48:30 device %busy avque r+w/s blks/s avwait avserv 07:48:31 fd0 0 0.0 0 0 0.0 0.0 nfs1 0 0.0 0 0 0.0 0.0 nfs219 0 0.0 0 0 0.0 0.0 sd1 0 0.0 0 0 0.0 0.0 sd1,a 0 0.0 0 0 0.0 0.0 sd1,b 0 0.0 0 0 0.0 0.0 sd1,c 0 0.0 0 0 0.0 0.0 sd1,g 0 0.0 0 0 0.0 0.0 sd3 0 0.0 0 0 0.0 0.0 sd3,a 0 0.0 0 0 0.0 0.0 sd3,b 0 0.0 0 0 0.0 0.0 sd3,c 0 0.0 0 0 0.0 0.0 sd6 0 0.0 0 0 0.0 0.0 Average fd0 0 0.0 0 0 0.0 0.0 nfs1 0 0.0 0 0 0.0 0.0 nfs219 0 0.0 0 0 0.0 0.0 sd1 0 0.0 0 0 0.0 0.0 sd1,a 0 0.0 0 0 0.0 0.0 sd1,b 0 0.0 0 0 0.0 0.0 sd1,c 0 0.0 0 0 0.0 0.0 sd1,g 0 0.0 0 0 0.0 0.0 sd3 0 0.0 0 0 0.0 0.0 sd3,a 0 0.0 0 0 0.0 0.0 sd3,b 0 0.0 0 0 0.0 0.0 sd3,c 0 0.0 0 0 0.0 0.0 sd6 0 0.0 0 0 0.0 0.0 07:48:30 rawch/s canch/s outch/s rcvin/s xmtin/s mdmin/s 07:48:31 0 0 147 0 0 0 Average 0 0 147 0 0 0 07:48:30 scall/s sread/s swrit/s fork/s exec/s rchar/s wchar/s 07:48:31 2637 0 95 15.79 15.79 0 19216 Average 2637 0 95 15.79 15.79 0 19216 07:48:30 swpin/s bswin/s swpot/s bswot/s pswch/s 07:48:31 0.00 0.0 0.00 0.0 116 Average 0.00 0.0 0.00 0.0 116 07:48:30 iget/s namei/s dirbk/s 07:48:31 0 195 121 Average 0 195 121 07:48:30 runq-sz %runocc swpq-sz %swpocc 07:48:31 2.0 526 Average 2.0 526 07:48:30 proc-sz ov inod-sz ov file-sz ov lock-sz 07:48:31 45/986 0 973/4508 0 357/357 0 0/0 07:48:30 msg/s sema/s 07:48:31 0.00 0.00 Average 0.00 0.00 07:48:30 atch/s pgin/s ppgin/s pflt/s vflt/s slock/s 07:48:31 0.00 0.00 0.00 505.26 1036.84 0.00 Average 0.00 0.00 0.00 505.26 1036.84 0.00 07:48:30 pgout/s ppgout/s pgfree/s pgscan/s %ufs_ipf 07:48:31 0.00 0.00 0.00 0.00 0.00 Average 0.00 0.00 0.00 0.00 0.00 07:48:30 freemem freeswap 07:48:31 15084 1224421 Average 15084 1224421 07:48:30 sml_mem alloc fail lg_mem alloc fail ovsz_alloc fail 07:48:31 2617344 1874368 0 17190912 10945416 0 3067904 0 Average 186953 133883 0 1227922 781815 0 219136 0 |