https://bugzilla.novell.com/show_bug.cgi?id=269741 gregkh@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|kernel- |pavel@novell.com |maintainers@forge.provo.nove| |ll.com | -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 pavel@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #3 from alpha096@tpg.com.au 2007-05-29 16:03 MST ------- This is very difficult in 2 sentences and being 100% re-producible I cannot understand the above however, 1. I normally use S2Disk. 2. If you reset the PC OR as in my case the complete boot process will not be observed as it appears that the last image file written to disk remain and the system will re-read it after it fixes the file system. 3. The problem is - please confirm that the Disk image file is deleted after resumption. It needs to be but appears not to be ever deleted just overwritten if S2Disk is used frequently. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #4 from pavel@novell.com 2007-05-30 05:02 MST ------- Are you seeing bootloader or kernel problem? Does it go away if you trigger swsusp manually? (echo disk > /sys/power/state)? Is the problem that if you do suspend to disk, then resume, then reset, it tries to resume from disk? Stefan, do we see something like that? I have bad problems understanding Scott's report. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #6 from seife@novell.com 2007-05-31 04:17 MST ------- (In reply to comment #5)

Let me try to assist.

Every time suspend to disk is used an image file is created

wrong.

- the problem is that the same image file does not appear to be deleted after successfully waking up - so it remains on the HDD.

also wrong. Which file do you think might be the image file on your disk? We suspend into the swap partition, so if you have a working swap after resetting your machine hard, the suspend image is deleted.

In the event of a system failure where a normal restart or shut-down has not occured the boot sequence appears to see the file left behind from when ever and reads it rather than going through a normal total boot procedure according

how do you get to this assumption? What happens on your machine that makes you think that it resumes instead of booting? Does it show the "Loading image data pages ... XX%" progress indicator?

to if the previous session is normally saved or he system is configured for a new session.

The problem is that the image file seems not to be deleted after successfully waking up.

There is no image file. Pavel: no, i did not see anything like that. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 seife@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |fseidel@novell.com, behlert@novell.com, | |hmacht@novell.com ------- Comment #9 from seife@novell.com 2007-06-01 01:16 MST ------- unfortunately comment #1 does not give enough detail. "is the same screen that would appear if the system had been shut down via suspend to disk" is not really a good description, since the same screen is shown during normal boot and during resume (unless you are canceling the bootsplash image by pressing "escape"). Can you name the "system analyst at suse.de"who can reproduce that, so that i can ask him to show it to me or give me access to his system? This would make debugging easier. I can not reproduce this. I use suspend to disk on my machine on a regular basis and my machine often crashes hard (due to a very experimental X driver :-), so i should, in theory, see the same problem as you do. But i don't. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #11 from alpha096@tpg.com.au 2007-06-01 22:04 MST ------- Created an attachment (id=143743) --> (https://bugzilla.novell.com/attachment.cgi?id=143743&action=view) self titled fie as requested -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #13 from alpha096@tpg.com.au 2007-06-01 22:06 MST ------- Created an attachment (id=143745) --> (https://bugzilla.novell.com/attachment.cgi?id=143745&action=view) self titled fie as requested -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #15 from alpha096@tpg.com.au 2007-06-01 22:50 MST ------- G'day frank. I am really very sorry cannot provide output of "/sbin/swapon -s" after such a resume?" Pulling the plug on a production system I fear. Yesterday I have a power outage and all PC recovered file journal's except 1 PC which holds the IDS Management system SQL database for many clients data. I lost over 120 Days of History and can only restore to last backup 7 days ago which I am still having trouble with indexing of the SQL database. Yes I know I should have a UPS. (And to cap the day off Novell called me from the US and told me I did not win the SLED Laptop) I have tried every means I know of capturing a screen image (SHFT+Print Screen), (ALT+Print Screen), CTRL+Print Screen) without any joy. I cannot even lock the image. Just to be really clear about the boot process difference I see but cannot capture. Normal boot after normal shut-down (1) 1. Loading Grub.... 2. Blue splash screen to select Kernel 3. Open Suse Splash screen. If you pull the plug on a PC that has had a previous Suspend to Disk I see.(2) 1. Loading Grub. 2. Text screen (the same screen you see in a normal resume) 3. Open Suse Splash screen 4. Further text screen 5. GUI KDE session as normal. I am also a little confused at comment #6 contradicting itself.

...

Every time suspend to disk is used an image file is created

wrong. We suspend into the swap partition, so if you have a working swap after resetting your machine hard, the suspend image is deleted. Anyway As has been stated by fears of the suspend to disk file(image) written to swap is deleted. Re-iterating comment #1Whilst the absence of the normal resume after shut down sequence is not present if you re-start after pulling the plug, it appears that there is little if no consequence apart from the user not being presented with the boot options screen as in example 1 item 2. The rest of the scenario I presented as possible is excluded as the image file is deleted so the rest of the scenario's logic is not founded. If there is a bug here its is only that if the user suspend and resume and the plug is pulled after such a resumption, the user misses out is the boot options selection screen. The subsequent session that is established after such an event is one of a normal O/S boot procedure including fschk.

-- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 seife@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |duwe@novell.com ------- Comment #17 from seife@novell.com 2007-06-04 01:49 MST ------- Ok, now i begin to understand. If you suspend, resume, then pull the plug, you still do not get the boot selection menu. Note that it is normal that for a resume you do not get that menu, since we set a flag in GRUB that tells it "please boot immediately THIS kernel, since we are going to resume from suspend". This flag should, however, be cleared after resume and the next reboot should again show the boot prompt. There might be interactions between journaling filesystems and hard reboots that might affect this, but it should only happen if you pull the plug very soon after resuming. Might i ask what filesystem is used for the /boot partition (or /, if /boot is not on a separate partition)? You can find out easily with "df -T" Adding our GRUB guru to CC:, maybe he has an idea what is happening. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 alpha096@tpg.com.au changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |seife@novell.com ------- Comment #19 from alpha096@tpg.com.au 2007-06-05 17:55 MST ------- RE comment #9 "I can not reproduce this. I use suspend to disk on my machine on a regular basis and my machine often crashes hard (due to a very experimental X driver :-), so i should, in theory, see the same problem as you do. But i don't." Now that you begin to understand you could please confirm that you are NOT ABLE to see the same difference in boot procedure as I have now described more clearly. Comment#17 the image (file) whatever you would like to call in written to the swap partition of the Disk "be cleared after resume. Please clarify "but it should only happen if you pull the plug very soon after resuming.". Is there a time delay feature set to delete at a specified time or is there a reliance on during normal operation /swap space is utilised by normal O/S application usage. I have noticed before hibernation, sometimes on my PC 1.5GB RAM /swap must be used as there is a text statement in the preparing for hibernation of ""Free Swap.....: OR Is this time delay a set parameter similar to a "flash dirty cache buffers" but in this case its "flush unneeded virtual memory" Please clarify also slightly as an aside : If Suspend to disk created an image in Virtual RAM why we have so many issues* with suspend to RAM and Suspend to Disk (RAM is RAM weather it be virtual or simulated. *Not dependent on Video issues? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #21 from seife@novell.com 2007-06-06 07:10 MST ------- (In reply to comment #19)

Now that you begin to understand you could please confirm that you are NOT ABLE to see the same difference in boot procedure as I have now described more clearly.

Yes, i am not able to see the same difference, and my machine crashes rather often due to me using an (very :-) experimental X driver.

Comment#17 the image (file) whatever you would like to call in written to the swap partition of the Disk "be cleared after resume.

Please clarify "but it should only happen if you pull the plug very soon after resuming.". Is there a time delay feature set to delete at a specified time or is there a reliance on during normal operation /swap space is utilised by normal O/S application usage. I have noticed before hibernation, sometimes on my PC 1.5GB RAM /swap must be used as there is a text statement in the preparing for hibernation of ""Free Swap.....:

OR

Is this time delay a set parameter similar to a "flash dirty cache buffers" but in this case its "flush unneeded virtual memory"

No, the reason i was thinking this could be a problem is, that on a journaling filesystem (like reiser), not everything that is written to disk is immediately written to the final place where it will end up. So what i was thinking was "we reset the preselected grub entry, but then the machine crashes, and the reset entry actually does not get written to disk. On the next boot, the journal is replayed and the old boot preference is restored" Torsten explained to me that this cannot happen in this case, since the bootloader restores the old default setting without going through the filesystem journal code, and without any caching effects etc. So this thought of mine was going into the wrong direction.

Please clarify also slightly as an aside :

If Suspend to disk created an image in Virtual RAM why we have so many issues*

It does not. It merely uses the same medium that is also used vor virtual memory (the swap partition) to store the data.

with suspend to RAM and Suspend to Disk (RAM is RAM weather it be virtual or simulated.

No, suspend to RAM is absolutely totally different from suspend to disk. They have basically nothing in common (apart from driver issues and stuff) Anyway, i'm getting the feeling that we won't be able to fix this problem without a setup where we can reproduce this nearby. So if you could name the suse.de guy, this would be very helpful. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 seife@novell.com changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |WORKSFORME ------- Comment #23 from seife@novell.com 2007-06-07 03:14 MST ------- I do believe that you see this bug, so it is not INVALID (invalid is somehow an alias for "reporter is doing something stupid" :-) just kidding) But i really cannot reproduce it, so i'll resolve it as WORKSFORME. I will carefully watch for such failures on my machines though, once i am able to reproduce it, it should be not too hard to track it down. Thanks for reporting and working with us on this difficult report. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.

https://bugzilla.novell.com/show_bug.cgi?id=269741 ------- Comment #25 from seife@novell.com 2007-06-08 07:01 MST ------- Don't get me wrong: i believe you that you are seeing this bug, and i will watch out carefully if i experience something similar here. However, right now i do not see anything like that and without having a machine here at hand on which i can reproduce this, i cannot really do much about it. It works for me, but i believe you that it does not work for you. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.