I will look into that. Unfortunately, the UNIX stations aren't close to my desk right now, but I should be able to get to them shortly. Do you think this is why I am getting the messages "alloc: /:file system full," and "realloccg: /:file system full" and "syslogd: /var/adm/messages: no space left on device"?
Well for that, we would need to know how many file systems you have and where there are mounted ( mount point ) :
/ is root, that is where /etc /dev reside
/var can be in / or as a separate file system...
(look at the df -k above..)
I'm a bit paranoid so when I configure machines I use file systems:
This is a HP-UX server (command bdf does not exist on solaris...).The reason is : Its easier to find some space to recreate a corrupted file sytem than a whole disk (and I keep aside enough space to be able to recreate the largest file system... that I can temporarily use to do some testing...)
I followed in 98 a course multiplatform UNIX (SUNOS, solaris 2.XX, AIX, DG-UX, IRIX) , and digging out a few notes I found:
solaris 2.X:
Use the command prtconf to see what hardware you have!
The devices are created in /devices and links are created in /dev to the device drivers...
(You have a format command: It is for formatting create partitions and check the state of a disk)
Convention
The whole disk can be partitioned in 8 maximum (0 to 7)
the slice 2 is for the whole disk
The information about the partitioning is kept in the label of VTOC
You can use the command prtvtoc to see the content
About /dev:
All devices drivers are in /devices. Because of SVR4 naming conventions the drivers are all referenced in /dev meaning you should not find one true file in /dev but links...
This will help us find the culprit if he is here:
explanation: long listing (ls -l) begins with file type and permissions 2nd position( but everything is a file under UNIX) and the perms are read write execute x 3 (owner, group, the rest...) if it start with anything else than - (or -r), it cannot be a (standard) file e.g l is for a link d is for a directory
Last edited by vbe; 10-20-2011 at 12:51 PM..
Reason: typos
So it looks like I was wrong. When I do du -sk/dev it shows 314. If I understand correctly, the -k means display in kb? In that case, I am far below 60 MB.
the results of df -k is this:
the other machine I am managing, which is up and running looks like this with th df -k
why does one have a problem and the other not, when both are full? There is a difference of 1.5MB more on the station that I can't login vs the one I can. I guess that can be the difference. You mentioed a couple of other things to look at, which I will try.
thanks.
---------- Post updated at 10:40 AM ---------- Previous update was at 09:51 AM ----------
trying the
shows
Is there something I should be looking for when doing the prtconf and prtvtoc? I entered the command prtconf and see devices, with one of three messages after: (driver not attached), instance #0, and insance #1. Prtvtoc shows
I am not sure what to get from that. I feel like such an unschooled newbie!
---------- Post updated at 11:11 AM ---------- Previous update was at 10:40 AM ----------
Another thing. Someone suggested doing the df -k, and finding the full disk, and fsck -y, which I did:
I have done the fsck -y command several times, and always after a reboot, it gives the file system state in superblock is wrong. It repairs it. I could type fsck -y again, and it is fine. I will reboot the system, and do fsck -y, and I will receive the same file system state in superbock is wrong, and it repairs it. Once I did not do the fsck -y, and when I went to boot up, I saw a superblock is wrong error. Could this be the cause or result of a problem?
Thanks,
Moderator's Comments:
Please use code tags also on generated output, data, logs etc, thanks.
Last edited by zaxxon; 10-20-2011 at 02:30 PM..
Reason: code tags
Sorry, I had to leave yesterday after a strong migraine access.. ( I usually stay when one is in need of help until bed time...)
I have no machines not LVM or equivalent so the commands I pass are not the same and I have to remember the older days...
Normally you have to boot in single user in order to use fsck of this command has to be passed file system unmounted, because the operating system buffers file system data...
I remember though on HP stations where there were only one file system on the disk there was a workaround, an fsck option and most important a reboot option (reboot -n cmd that was...) to not sync the file systems before shutdown, for a clean shutdown waits for the buffer to be flushed...
I'm surprised by the difference of size of /usr which makes me wonder if there were not a dirty trick with links... I would suspect links missing - or existing on the box from / to /usr, I'm thinking of links like /tmp ->/usr/tmp or /var/tmp -> /usr/tmp ( but /tmp would be tmpfs from swap...)
Then most important, when a system crashes it produces a coredump or core file, in your situation I would search for them and get them removed, starting by /
If you see one there remove it!
Then we can go for some cleaning:
-i for interactive, in those days you wouldn't bother but now there are so many packages with directories called myproduct/mystuff/core... better be safe
Does that help?
And yes zaxxon in right you must use code tags for your data and code:
vbe, thanks for your concern and your help. I can understand why you would have a migraine after all my questions.
I have tried using code tags, but some of previous posts have come from smart phone use, which is what I had available at the time, which in itself is somewhat cumbersome, and using the rest of the features difficults as well. I will try to be better. I am not used to forum use, and so I apologize for for that.
Is there a difference between single user and maintenance mode on the machines? I am assuming that they are the same, but I have heard of both being used. I have used boot -s, after it has booted up mentions system maintenance, or something like that. Is this the same as single user? This is the only way I have been able to login on the machine, entering the root password. Then I can do things. This is the only way I have been able to use fsck.
I will start looking for links in /, and use the find command. I will let you know what I find out.
This next question shows my forum ignorance, but that big "thanks" button at the bottom corner of each post. Is that for each time I feel you have helped me learn something new? Or would that be used for the final fix of the problem? I would like to give you credit for all the help you have given me, but want to do it at the appropriate place.
I have the input file like this.
Input file: 12.txt
1) There are one or more than one <tr> tags in same line.
2) Some tr tags may have one <td> or more tna one <td> tags within it.
3) Few <td> tags having "<td> </td>". Few having more than one " " entry in it.
<tr> some td... (4 Replies)
Hi All,
Please help me and guide me to write a bash/shell script on Linux box to delete parent entry with all their child entries.
example:
Parent is :
----------
dn: email=yogesh.kumar@wipro.com, o=wipro, o=in
child is:
----------
dn: cn: yogesh kumar, email=yogesh.kumar@wipro.com,... (1 Reply)
1.) I am to write scripts that will be phasetest folder in the home directory.
2.) The folder should have a set-up,phase and display files
I have written a small script which i used to check for the existing users and their password.
What I need help with:
I have a set of questions in a... (19 Replies)
can anyone explains me the last fields in the below cron job. Here it will run on 31st 23:59.. what is mean by 1-7/2 (first to seventh month) what /2 represents
59 23 31 1-7/2,8-12/2
Thanks in advance. (2 Replies)
Hi,
i need to setup a cronjob that has will execute iostat command from morning to evening time.
for instance the timing has will be like this.
8:00 A.M -- 6:00P.M
how to define this entry in crontab
Regards (3 Replies)
I had a drive go bad. I have replaced the drive and have my system up. I have a 1-to-1 mirror raid. My mirrored boot drive brings the system up. I issued a reboot command. The system comes up gives me a login prompt and generates the error: failed write of utmpx entry:"i2".
What is i2? (1 Reply)
Want to make a entry for using shell script, something like this:
Name:_____________ Age:____________
address:___________
This entry form showed on the screen when the script executed, ask the user enter the field one by one ( user press <Enter>, cursor will be on the next field), anyone... (5 Replies)
When shutting down a freshly installed version of 5.7 on a Compaq server I get two messages:
INIT: failed to write of utmpx entry: "s0"
and:
INIT: failed to write of utmpx entry: "fw"
then I can continue with the shutdown. When I bring the system back up it has not saved any of my... (2 Replies)
:confused:
HI...
I am a brasilian software maker.
I have a message " utmpx - filesystem isfull - /var/messages/unable to edit data "
Like resolves its problem ???
Tank yours.... (2 Replies)
I am trying to set up a trust between two hosts and from another host to both by creating an /etc/hosts.equiv file on both. I am able to rsh to all but one host (and I am able to rsh from that host to the other). I receive the following error:
No utmpx entry. You must exec "login" from the lowest... (1 Reply)